Jonathan Loh & David Harmon
This research and report has been commissioned by WWF Netherlands
About the Authors
Jonathan Loh is a biologist specializing in monitoring, assessment and
indicators of biological and cultural diversity. He works for several
conservation NGOs and is an Honorary Research Associate of the
Zoological Society of London.
David Harmon is executive director of the George Wright Society,
a professional association of people who work in parks, protected areas,
and cultural sites. He also maintains an active research interest in the
relationship between biological and cultural diversity, having co-founded
the NGO Terralingua, which is devoted to that subject.
We would like to thank Johan van de Gronden, Monique Grooten,
Natasja Oerlemans and Natascha Zwaal of WWF-Netherlands for their
enthusiasm, support and comments on earlier drafts of this report.
We are also grateful to Terralingua and The Christensen Fund for
supporting precursors to the work presented here. We should point out
that the opinions expressed in this report are those of the authors and not
necessarily those of WWF or any other of the organizations we work with.
WWF is one of the world’s largest and most experienced independent
conservation organizations, with over 5 million supporters and a
global network active in more than 100 countries.
WWF’s mission is to stop the degradation of the planet’s natural
environment and to build a future in which humans live in harmony with
nature, by conserving the world’s biological diversity, ensuring that the
use of renewable natural resources is sustainable, and promoting the
reduction of pollution and wasteful consumption.
Published in June 2014 by WWF–Netherlands, in association with
authors. Any reproduction in full or in part of this publication must
mention the title and credit the above-mentioned publisher as
the copyright owner.
The designation of geographical entities in this report, and the
presentation of the material, do not imply the expression of any opinion
whatsoever on the part of WWF or the authors concerning the legal status
of any country, territory, or area, or of its authorities, or concerning the
delimitation of its frontiers or boundaries.
Loh, J. & D. Harmon. 2014. Biocultural Diversity: threatened species,
endangered languages. WWF Netherlands, Zeist, The Netherlands.
peer&dedigitalesupermarkt (www.pdds.nl) and Jonathan Loh
Front cover photograph: Hadza boy, speaker of a Khoi-san language,
hunting. Fewer than 1,000 speakers of Hadza remain. Lake Eyasi,
Tanzania © Jonathan Loh
Biocultural Diversity 2
Evolution of Species and Languages 6
Decline of Biocultural Diversity 22
Status of Species and Languages 30
End Notes 52
Data Tables 53
Nature and culture as dual
aspects of a single entity
The Oxford English Dictionary oers the following denitions (OED Online):
The phenomena of the physical world collectively; esp. plants, animals, and other
features and products of the earth itself, as opposed to humans and human creations.
The distinctive ideas, customs, social behaviour, products, or way of life of a particu-
lar nation, society, people, or period. Hence: a society or group characterized by such
It is customary to think of nature and culture as being quite dierent, belonging to en-
tirely separate domains, one contains items such as butteries, the Amazon rainforest
and photosynthesis, while the other contains items such as Beethoven’s piano sonatas,
wedding ceremonies or sushi. Yet nature and culture often interpenetrate and overlap.
What is wine-making, bee-keeping or gardening: nature or culture? They are undoub-
tedly human activities, and each has its own culture, but there is a strong element of
nature involved. What about varieties of domesticated plants and animals? They are
human creations because their genomes have been altered by thousands of generations
of selective breeding, and particular breeds may be associated with particular places
or peoples, so they are as much a product of culture as of nature. What about lands-
capes? Is there anywhere left in the world that is entirely natural, untouched by human
intervention? The deep sea bed perhaps, and possibly Antarctica; but most landscapes
are, to a greater or lesser extent, the product of human culture too. Even the Amazon
rainforest is what it is not just because of the natural evolution of its ecosystems, but
also because of centuries of human manipulations to those ecosystems. So would it
make more sense to think of all the myriad manifestations of nature and culture as
expressions of a single concept, a nature-culture nexus?
We can think of nature and culture as being dual aspects of a single entity,
biocultural diversity; but not just because the two concepts are blurred at their
interface. It is because both nature and culture, as dened above, are what they are as
a result of evolution, and they have evolved in similar ways. So similar, in fact, that
in this report we will describe culture and cultural evolution in the same terms as
nature and natural evolution, using concepts borrowed from genetics, ecology and
population biology. We will go on to examine the extinction crisis facing both biologi-
cal and cultural diversity, and use methods developed in conservation biology to assess
and compare the state of biodiversity with the state of cultural diversity, and contrast
recent trends in the two.
In order to assess status and trends we need a unit of measure. Biodiversity and
cultural diversity are such broad concepts that we need to focus on something specic
BOTH NATURE AND
CULTURE ARE WHAT
THEY ARE AS A RESULT
OF EVOLUTION, AND
THEY HAVE EVOLVED IN
and measureable, so we have chosen two fundamental units or classiers of nature and
culture: species and languages. Species are the basic units of biodiversity; languages
are a useful proxy to stand for the world’s diverse cultures. Other elements of biodiver-
sity such as ecosystems or genes, and other aspects of culture such as religions, arts, or
livelihood and subsistence strategies, are much harder to dene and very much harder
There are striking parallels between species and languages (Harmon 2002). A species
is a group of similar individual organisms that is capable of interbreeding. The ability
to produce fertile ospring is fundamental to the biological denition of a species.
Horses and donkeys belong to dierent species, even though they are closely related,
as their ospring, mules, are infertile. Humans all belong to a single species, Homo
sapiens. The genetic variation among humans is remarkably small, reecting the fact
that the modern human species is relatively young, only about 200,000 years old, and
yet there is a staggering amount of cultural and linguistic variation among the human
population (Pagel & Mace 2004). Linguists identify around 7,000 languages spoken
worldwide (Lewis et al. 2013). By analogy with the denition of a species, two human
individuals can be said to speak the same language if they can understand one another.
If they nd each other unintelligible, they are speaking dierent languages.1 Dialects,
by this denition, are analogous to subspecies: communication is possible between
two individuals, although it may not be as easy. There are several subspecies of tiger,
Siberian, Bengal or Sumatran for example, which can interbreed successfully in zoos,
but their geographic ranges do not overlap in the wild. Given time, sadly something
which is not on the tiger’s side, the geographically isolated subspecies would evolve
into reproductively separate species, a process known as speciation. New languages
can evolve through a process that is akin to biological speciation, and the formation of
dialects is the rst step along the path to the evolution of two separate languages, pro-
vided that there is limited intercommunication between the two dialect populations.2
By using species to stand for all biological diversity and languages to stand for cultural
diversity we are taking a narrow view, but making a useful simplication at the same
time. Biological diversity is broader than species richness. It spans across scales from
genes and proteins at the microscopic level to ecosystems and landscapes at the ma-
croscopic level. Species lie somewhere in the middle, but as the carriers of genes and
the components of ecosystems, they can fairly represent all biological diversity. In the
same way, languages will stand as a proxy for all of cultural diversity, from the micro
level of words, ideas and behaviours to the macro level of peoples and societies.
Fisherman, a speaker of a Trans-New Guinean language, hanging
nets up to dry. Western Province, Papua New Guinea.
© Brent Stirton / Getty Images / WWF
LANGUAGES Explaining biocultural diversity
in terms of the Tree of Life
The parallels between species and languages have been noted and commented upon
since the 19th century, famously by Charles Darwin in The Descent of Man (1874).
The formation of dierent languages and of distinct species, and the proofs that
both have been developed through a gradual process, are curiously parallel…. We
nd in distinct languages striking homologies due to community of descent, and
analogies due to a similar process of formation…. The frequent presence of rudi-
ments, both in languages and species, is still more remarkable….In the spelling
also of words, letters often remain as the rudiments of ancient forms of pronun-
ciation. Languages, like organic beings, can be classed in groups under groups;
and they can be classed either naturally according to descent, or articially by
other characters. Dominant languages and dialects spread widely, and lead to the
gradual extinction of other tongues. A language, like a species, when once extinct,
never, as Sir C. Lyell remarks, reappears.
To illustrate the analogy between species and languages, picture that well-known
Darwinian metaphor, the Tree of Life. The biocultural version of the tree diers from
the usual version in that it has gone through not one but two distinct types of bran-
ching or diversication; the second diversication took place near the end of one the
myriad outer twigs of the rst tree (Figure 1). The rst diversication was the evolution
of multicellular organisms on Earth today, and the second diversication represents
the evolution of human cultural diversity. Both of these evolutionary diversications
can be represented as trees, or phylogenies, but with one tree growing from the end of
one branch of the other.3 Figure 1 shows these two great radiations – the biological and
the cultural – on a log scale. The rst radiation took place near the bottom of the tree,
between around 550 million years ago and the second radiation occurred about half-
way up the tree at around 70-80 thousand years ago.
Life rst appeared around 3.5-4.0 billion years ago; the earliest fossils of the simplest
cells, bacteria, date back about 3.5 billion years. More complex life did not evolve until
1.8 billion years ago when the rst plant and animal cells appeared in the form of algae
and protozoa. These eukaryotic cells contained a nucleus to hold their DNA and had a
more complex internal structure, which arose from the symbiotic union of prokaryotic
(bacterial) cells, but remained unicellular. For the rst three billion years of the evolu-
tion of life on earth the most complex organisms were single-celled organisms. It was
not until about 550 million years ago that colonies of cells grouped together into the
rst multicellular life forms, known as the Ediacaran fauna, which resembled quilted
discs and pillows.
TREE DIFFERS FROM
THE USUAL VERSION
IN THAT IT HAS GONE
THROUGH NOT ONE BUT
TWO DISTINCT TYPES
The Ediacaran period lasted only a few millions of years before the quilted pillows were
blown away in a massive, unprecedented and unrepeated diversication of animal
biota that happened around 540 million years ago, known as the Cambrian Explosion.
This explosion, or radiation, produced new life forms or species more rapidly than at
any time before or since. Multicellular organisms appeared of enormous complexity
by comparison with the Ediacaran fauna, including some of the most bizarre animals
in the fossil record: many had hard body armour and possessed a range of formida-
ble weaponry. Within a geological blink it was all over, but the Cambrian Explosion
had produced myriad life forms including all known basic body plans of animals. The
ancestors of arthropods, molluscs, annelid worms, echinoderms and all other modern
phyla including the chordates (and therefore us) were there in one form or another,
and all animal species since that time have conformed to the basic blue prints that
evolved in that sudden burst of activity.
The reason for the Cambrian Explosion is unknown, but a number of possibilities
have been proposed. The entire planet during the time immediately preceding
the Cambrian was glaciated, a period known as Snowball Earth (Walker 2003).
The warming that ended Snowball Earth seems to have jump-started a new phase of
multicellular evolution: initially the Ediacaran, followed by the Cambrian Explosion.
Time (log scale)
540 million years ago:
70,000 years ago:
out of Africa
3.9 billion years ago:
life on Earth begins
Tree of Life
The biological tree (in green)
of species diversity began
its diversication with the
Cambrian Explosion around
540 million years ago;
the cultural tree (in red) of
linguistic diversity began to
diversify about 70-80,000
years ago, near the end of
one of the myriad branches
of the biological tree.
Another physical change at that time in Earth’s history was a rise in the atmospheric
oxygen content to its current level of around 21%, which would have aided the evolu-
tion of complex multicellular organisms. A third possibility is that a new type of gene
that controls morphological development in the embryo, known as Hox genes, rst
appeared at the time of the Cambrian Explosion, enabling a plethora of new body plans
The Cultural Explosion
The Tree of Life continued to branch and grow, continually evolving new species and
losing old ones through extinction. Where whole groups of species died out, those
branches stopped growing. This process went on for more than half a billion years,
until the number of individual twigs at the outer edge of the tree numbered in the mil-
lions. Then an extraordinary, unparalleled event occurred at the end of one of its twigs.
To an external observer, that twig would not have appeared exceptional, for although
it represented a large mammalian species, it was by no means the biggest, or fastest,
or the one with the most impressive body armour or weapons. But at some point, for
reasons that are still unknown, the species on that twig began to talk. That species was
our own, and as a result of our remarkable and unique innovation, language, the tree
began a second massive evolutionary radiation, as signicant as the Cambrian Explo-
sion 540 million years earlier.4
Modern Homo sapiens rst appeared only around 200,000 years ago, but we can trace
our lineage back to the last common ancestor that we share with our closest living
cousins, the chimpanzee and the bonobo, who lived about six million years ago. Exactly
how or when language evolved is not known. But once it had taken hold it enabled an
entirely new mode of evolution to take o – cultural evolution. Cultures evolve like
species in many ways. Cultural items or traits are subject to hereditary transmission,
variation by mutation and selection: the prerequisites of evolutionary change.
Heredity in biology involves passing genetic information encoded in DNA from parent
to ospring. The hereditary transmission of culture is mediated not by passing DNA
from parent to ospring, but by one individual learning something from another, be it
an idea, a behaviour, custom or another aspect of a way of life, and this transmission
is greatly facilitated and accelerated by means of language. One can think of cultural
information being transmitted as memes – the cultural analogue of genes (Dawkins
2006, Dennett 2002. See box “What is a Meme?” for further explanation). A meme,
such as a song for example, existing in the brain of one individual is passed out of the
mouth and via the ear into the brain of another individual. The meme has been copied,
and can be replicated again and again in the brains of more individuals. Variation
among memes occurs in a manner similar to mutation in genes. The tune of the song
can be altered, the words can change, verses added or dropped. Selection is carried
out by the individuals who come in contact with the meme, sometimes deliberately,
sometimes unconsciously. A memorable, useful or otherwise interesting meme will
be replicated many times, and spread successfully through a population – it becomes
an element of a culture. Less memorable memes will be less successful; unmemorable
memes will be forgotten.
CULTURE IS MEDIATED
NOT BY PASSING DNA
FROM PARENT TO
OFFSPRING, BUT BY ONE
ANOTHER, AND THIS
AND ACCELERATED BY
MEANS OF LANGUAGE
Ashaninka woman, a speaker of an Arawakan language. Nueva Victoria, Yurua River, Ucayali Province, Peru.
© James Frankham / WWF-Canon
What is a Meme?
The word meme was coined by Richard Dawkins in his book The Selsh Gene (2006),
and the word itself has become a successful meme. Dawkins proposed memes as the
basic units of cultural evolution, and the idea has been developed by other thinkers
such as Daniel Dennett (2002), although it remains controversial and is not widely
accepted by sociologists and cultural theorists. Dawkins introduces the concept of
cultural evolution with a linguistic example:
Georey Chaucer [c.1343-1400] could not hold a conversation with a modern
Englishman, even though they are linked to each other by an unbroken chain
of some twenty generations of Englishmen, each of whom could speak to his
immediate neighbours in the chain as a son speaks to his father. Language
seems to evolve by non-genetic means, and at a rate which is orders of
magnitude faster than genetic evolution.
Just three conditions are necessary and sucient for evolution to occur: replication
(or heredity), variation (or mutation) and competition (or selection). If these three
conditions are met, evolution will happen. In nature, the rst two conditions are met
by DNA, the molecule that encodes genetic information in all plants and animals,
which replicates itself, but not perfectly (because of mutation). The third condition is
provided by the fact that resources are nite, so individual organisms (and therefore
their DNA) must compete for them in the Darwinian struggle for existence. But is there
any other material apart from DNA which can replicate with variation and competes in
the struggle for life? Yes, but not a material in the literal sense. Ideas. Memes are ideas
which meet the three conditions. Firstly, an idea can be copied from the brain of one
individual to the brain of another, so it can replicate. Secondly, ideas show variation
from one individual brain to another, either because of imperfect copying or because
of an innovation by an individual brain. Finally, ideas compete for nite brain-space
in a population of brains, some are very successful and become very common, others
are less successful and remain rare, and others still are not copied at all and go extinct.
These ideas could be songs, stories, games, recipes, customs, clothing, art, technolo-
gies, anything in fact that constitutes culture. The basic unit of cultural heredity is the
meme. The song “Happy Birthday to You” is an example of a phenomenally successful
meme. As is wearing a tie. How to make an origami frog is a less successful meme, as it
is more dicult to learn, and not very useful, but that also makes it a more interesting
meme. A language is not a meme, it is a vast collection of memes working together – a
meme complex – but which are largely copied as a group. Most elements of culture
are in fact meme complexes. The rules of chess, for example, is a very successful set of
memes, although not every brain remembers them all correctly. There are many other
chess memes, such as opening gambits which are very successful at propagating
themselves, but only among serious chess players.
It takes several years for the infant brain to acquire language, but once it has grasped
it suciently, it enormously accelerates the learning of other types of behaviour, such
as how to make tools, to hunt animals and gather edible plants, to cook, to grow crops
and raise livestock, to make clothes, to build homes, to tell stories, to paint pictures, to
write books, to play games, to do anything, in fact, that constitutes culture. Language
is a tool for encoding and transmitting memes. Because of the fundamental importance
of language to culture, linguistic evolution usually goes hand-in-hand with cultural
evolution, and languages can be viewed as a proxy for cultural groupings in terms of
the Tree of Life. It is possible for cultural behaviour to be transmitted between linguis-
tic groups, but the transmission is faster and more accurate within a single group.
The 7,000 languages spoken in the world today represent the outermost twigs of the
second tree, but of course there are many more extinct languages whose branches
ended before reaching the outer edge of the tree. Like species, some languages can
be classied into closely-related families, while others stand alone in families of one.
Languages belonging to the same family have a common ancestor, just as families of
species do. Germanic languages for example must have evolved from a single, ancestral
proto-Germanic language. All Germanic languages belong to the much larger, exten-
ded family of Indo-European languages, which includes among others French, Irish,
Greek, Russian, Persian and Hindi. Indo-European includes about 430 languages and
is among the half-dozen largest families, of which the most diverse are Austronesian
(about 1,250 languages spoken in Southeast Asia and the Pacic) and Niger-Congo
(about 1,500 African languages) (Lewis et al. 2013).
Perhaps the biggest dierence between biological evolution and linguistic-cultural evo-
lution is speed. Biological evolution is slow while cultural evolution is so rapid that it
can be observed taking place within our own lifetimes or even in front of our very eyes
and ears: by watching a lm made more than 50 or 60 years ago it is possible to hear
how much language has changed in terms of pronunciation, accent and some words
and phrases. Hence classifying languages into evolutionary families is tricky, as many
if not all similarities between related languages can be erased within a few centuries.
This is the problem that bugs the reconstruction of ancestral languages such as proto-
Indo-European, and makes it paradoxically harder for linguists to draw phylogenetic
trees and date the appearance and disappearance of languages than it is for biologists
to draw and date phylogenetic trees of species. Another major dierence between
species and languages is that borrowing occurs far more readily between languages.
Borrowing words is the equivalent of dierent species exchanging genetic material,
something bacteria can do easily, but is less common among multicellular organisms.
English is an example of a language whose origins lie in one part of the tree, the Ger-
manic branch, but has incorporated a vast number of words from a language belonging
to a neighbouring branch, French.5 So horizontal transmission of language, and hence
culture, does take place, but not so much as to destroy the basically tree-like structure
of cultural diversity.
Walking the talk
Around 100,000 years ago the modern human population comprised somewhere in
the order of 100,000 individuals, largely conned to the African continent, with a
few living north of the Sahara and as far east as Palestine. Between 70-80,000 years
ago, during the last ice age, people began to migrate out of Africa, probably crossing
from the Red Sea to the Arabian Peninsula as sea levels were much lower, rather than
moving north across the Middle East where deserts barred the way.6 From Arabia,
the migrants spread inexorably across Asia, probably following coastlines and moving
up river valleys. Their descendants colonised South Asia rst, and reached East Asia
60-70,000 years ago. About 40-60,000 years ago they succeeded in crossing the straits
between mainland Southeast Asia and Australia.
One pathway led the migration north from the Arabian Sea up the Persian Gulf, at that
time a broad, forested river valley into which the Tigris and Euphrates owed. This was
the route that led to Mesopotamia, Anatolia and eventually, around 30,000 years ago,
to Europe. Another pathway led into northern Asia, and then, about 15,000 years ago,
across the Bering Sea into North America. Within another thousand years the descen-
dants of the rst migrants to America reached Tierra del Fuego.
LIKE SPECIES, SOME
LANGUAGES CAN BE
OTHERS STAND ALONE
IN FAMILIES OF ONE
TO THE SAME
FAMILY HAVE A
JUST AS FAMILIES OF
The last great migration was not by land but by sea, across the Pacic Ocean from
Southeast Asia around 5,000 years ago, nally reaching New Zealand about
1,000 years ago (and, surprisingly, eastward across the Indian Ocean to reach
Madagascar about 1,200 years ago). The human colonization of the globe, save for
Antarctica, was complete.
Between 160,000 and 80,000 years ago, back in Africa, humans learned to make
composite tools such as spears, decorate themselves with beads, catch seafood, use
pigments and carry out ritual burials. It is possible that this was the period in which
complex language, culture and art rst appeared,7 and people carried these with them
as they crossed the globe. As they spread, living in small isolated groups, cultural and
linguistic evolution would have rapidly given rise to thousands of local and regional
variations, leading ultimately to a vast diversity of human languages and cultures.
Drawing the Family Tree
One way to look at linguistic evolution in action is to compare closely related modern
languages. For example, English and Dutch are related languages in the Germanic
family, descended from a common ancestral language that would have been spoken
somewhere on the northwest coast of mainland Europe around 2,000 years ago.
Frisian is a language that is spoken in Friesland in the north of the Netherlands8 and is
closely related to both Dutch and English. Scots is a language also very closely related
to English that is still spoken at home by a substantial proportion of the population in
Scotland; it contains many words that are closer to their Germanic roots than modern
English, and would be unintelligible to most English speakers outside Scotland. It is
easy to see how the languages have diverged from a common root by comparing words
for the numbers from one to ten (Table 1).
Number Dutch Frisian Scots English
1een ien ane one
2twee twa twa two
3drie trije thrie three
4vier fjouwer fower four
5vijf ﬁif fyve ﬁve
6zes seis sax six
7zeven sân seiven seven
8acht acht aicht eight
9negen njoggen nyne nine
10 tien tsien ten ten
The similarity between English and Frisian is demonstrated by the saying “Good butter
and good cheese is good English and good Fries”, which when spoken sounds virtually
the same in both languages. Tellingly, the saying also demonstrates the importance of
dairy-based agriculture in both cultures. So it is not dicult to imagine members of
the Frisii tribe after the end of the western Roman Empire in 410 CE (along with their
cousins the Anglii and the Saxones) crossing the North Sea to Britain, at rst as raiders
but later as settlers, taking with them their language and perhaps a few of their cows.
The names of
numbers 1-10 in four
Baka woman, a speaker of a Niger-Congo language, with collecting basket, gathering plants in the forest of La trinationale de la Sangha.
Central African Republic.
© Martin Harvey / WWF-Canon
Linguists have made comparisons between the thousands of languages spoken in
the world in order to work out the evolutionary relationships between them. These
comparisons rely on similarities between words which have descended from common
ancestral words, like the numbers one to ten in Dutch, English, Frisian and Scots. Such
words are termed cognate, and by knowing how sounds have changed systematically
in dierent languages over time, comparative linguists have reconstructed language
phylogenies or family trees.
European visitors to India as early as the sixteenth century began to notice similarities
between Sanskrit, Latin and Greek, but the most famous of these was William Jones
(1746-94), who is considered to be the founding father of comparative linguistics.
Jones was a scholar and magistrate living in Calcutta in the 1780s. He was a polyglot
uent in a dozen languages and familiar with two dozen more. He became fascinated
with Indian culture and co-founded the Asiatic Society of Bengal in 1784. In a paper he
delivered to the Asiatic Society in 1786 he noted that Sanskrit, Latin and Greek bear
…a stronger anity, both in the roots of verbs and the forms of grammar, than
could possibly have been produced by accident; so strong indeed, that no philo-
loger could examine them all three, without believing them to have sprung from
some common source, which, perhaps, no longer exists... there is a similar reason,
though not quite so forcible, for supposing that both the Gothic and the Celtic…
and the old Persian might be added to the same family (Jones 1824).
That common source came to be known as Proto-Indo-European (or sometimes
proto-Indo-Germanic) or PIE. It is not dicult – with hindsight - to see how Jones and
others reached their conclusion. Look, for example, at the names of the numbers one
to ten in Sanskrit, Latin, Ancient Greek, Gothic (an old Germanic language), Welsh
Number Ancient Greek
1oinos una eka Ek ains un
2duo duo dvi do twai/twos/twa dau/dwy
3treis tres tri tin þreis tri/tair
4tessares quattuor chatur car ﬁdwor pedwar/pedair
5pente quinque pancan panch ﬁmf pump
6hex sex sash chhah saíhs chwech
7hepta septem saptan sat sibun saith
8okto octo ashta Ath ahtau wyth
9ennea novem navan nau niun naw
10 deka decem dasan das taíhun deg
By using this method a family tree for the Indo-European languages was being put
together even before Darwin proposed his theory of evolution of species by descent
from a common ancestor (see Figure 2).
The names of
numbers 1-10 in
Language Family Tree
(after Gray et al. 2011)
Proto-Indo-European is the
inferred or reconstructed
ancestral language from
which all Indo-European
languages stem. It would
have been spoken about
9,000 years ago in Anatolia
(modern Turkey), the
language of the Neolithic
revolution and the rst
spread into Europe and Asia
carried by the expansion
of agriculture and evolved
into the hundreds of modern
(as well as thousands of
So who were the original Proto-Indo-Europeans and where did they come from? One
of the most surprising ndings of the comparative linguists was that some extinct lan-
guages spoken in Bronze Age Anatolia (modern Turkey) such as Hittite belonged to the
Indo-European family, even though modern Turkish, a member of the Altaic language
family, does not. The Hittite language was known only from cuneiform inscriptions
on clay tablets dating from the second millennium BCE and deciphered by the Czech
linguist Bedrich Hrozný in the early 20th century (Hrozný 1917). The key breakthrough
came when Hrozný found the word watar in a sentence alongside an ideogram, or
symbol, known to mean ‘bread’ in Sumerian. The similarity of watar to water, or Was-
ser in German, and the similarities of another word in the sentence to eat or essen,
and of another to aqua in Latin, led him to guess the meaning was something like eat
bread, drink water. But the shock was that this nding placed Hittite in the Indo-Euro-
pean family, not with the Altaic family of central Asian languages, or the Afro-Asiatic
family along with other Middle Eastern languages such as Arabic (see gure 9).
A great many theories have been proposed to locate the homeland and time of the
Proto-Indo-Europeans. Recently, scientists have re-examined the linguistic evidence
using computational methods developed to determine the evolutionary relationships
between species based on their DNA. Instead of comparing DNA, Gray and Atkinson
(2003) compared a list of 200 words in 87 languages to reconstruct the Indo-European
family tree (see Figure 2). They were able to date its origin to about 8-9,000 years ago
(Gray and Atkinson 2003, Gray et al. 2011), supporting the theory proposed by the
archaeologist Colin Renfrew that the Indo-European languages were carried from
Anatolia into Europe and South Asia on a cultural tsunami caused by the greatest
seismic event in prehistory – the Neolithic revolution, or the adoption and spread of
agriculture (Renfrew 1987).
A number of comparative linguists such as Joseph Greenberg have attempted to
construct higher-order language families, which unite several families into a single
grouping, equivalent to a class or phylum in zoology (Greenberg 2000). Indo-
European has been combined with several other families including Uralic (which
covers Finland and western Russian Arctic), Altaic (Siberia, Mongolia, Central Asia
and Turkey) and Eskimo-Aleut (North American Arctic) languages into a phylum
called Nostratic or Eurasiatic. Greenberg also proposed a super-family called Amerind
to include all but two indigenous American languages families. These proposals are not
widely accepted among linguists, but there is some support for Eurasiatic, based on
quantitative techniques used by evolutionary biologists, who estimate the date of the
proposed ancestral language, proto-Eurasiatic, would have been around 15,000 years
ago, close to the end of the last ice age (Pagel et al. 2013).
Some experts have even attempted to link languages as far apart geographically as
Basque and Navajo, along with a few northern Caucasian and Siberian languages, into a
phylum called Dene-Caucasian.9 Such heroic attempts at reconstructing deep historical
links between languages are highly controversial among linguists. If such a language
as proto Dene-Caucasian ever existed, or more likely but still highly controversially,
if not a language then at least some proto-Dene-Caucasian words existed, they would
have been spoken right back at the height of the last ice age, by Palaeolithic hunter-
gatherers living alongside mammoths and woolly rhinoceroses somewhere in Siberia.
However remote this possibility may sound, there is some genetic evidence that could
support this theory, from the 24,000-year-old remains of a young boy buried under
a stone slab in the village of Mal’ta near Lake Baikal (Raghavan et al. 2014). Analysis
of his genome revealed European ancestry which suggests that there had been an
eastward migration of people into Siberia from Europe.
But more surprisingly, his DNA showed he is also ancestral to modern Native Ame-
ricans, which suggest that some descendants of the Mal’ta population interbred with
East Asians in Siberia who then migrated across the Beringian land bridge to the
Americas around 15,000 years ago. Is it possible that the Basque and Navajo languages
retain some residual cultural imprint from those times?
Basque is one of the few remaining languages left in Europe that is not Indo-European
in origin. Genetically, the Basque people also show some dierences from other Eu-
ropean populations (Cavalli-Sforza 2000). Interestingly, places with Basque-derived
names and sites of Palaeolithic cave art overlap geographically in southwest France
and northern Spain. The earliest cave paintings date back 30-40,000 years, marking
the beginning of western art, and coinciding with the arrival of modern Homo sapiens
in Europe (long after the arrival of the Neanderthals, who soon disappeared). The cave
artists continued to produce their work for another 20,000 years, achieving their
magnum opus in the extraordinary paintings of horses, bison, aurochs, reindeer, as
well as more abstract images of humans, at Lascaux in France and Altamira in Spain,
around the time of the last glacial maximum around 17-18,000 years ago. Is it possible
that these artists spoke a language that was the ancestor of Basque? Indo-European
languages spread into Europe from Anatolia alongside the adoption of agriculture, as
settled farming replaced nomadic hunter-gathering as the primary way of life, and the
original languages spoken by the rst modern humans in Europe fell like dominoes.
Basque, for reasons that are unknown, is the last domino standing, a language isolate
descended from Palaeolithic hunter-gatherer-artists.
SOME LINGUISTS HAVE ATTEMPTED TO CONSTRUCT HIGHER-ORDER LANGUAGE
FAMILIES, WHICH UNITE SEVERAL FAMILIES INTO A SINGLE GROUPING,
EQUIVALENT TO A CLASS OR PHYLUM IN ZOOLOGY
Map 1: Global
(Stepp et al. 2004)
The diversity of languages
(black dots) strongly
correlates to areas of high
plant diversity (darker
A big coincidence
There is another way in which the evolution of languages mirrors the evolution of
species: the similarity in the geographic distributions of languages and of species
around the world. Places with high species diversity, especially tropical forests,
tend to have high linguistic diversity, and areas of low species diversity, such as tundra
and deserts, have low linguistic diversity (Mace & Pagel 1995, Nettle & Romaine 2000,
Moore et al. 2002, Sutherland 2003, Stepp et al. 2004, Loh & Harmon 2005).
The island of New Guinea which makes up less than one percent of the Earth’s
habitable surface, apart from being one of world’s biodiversity hotspots with
endemic species such as birds of paradise and tree kangaroos, supports around
1,000 languages, one seventh of the total. A glance at Map 1 conrms this view,
and that it is not just that places with greater population density have greater
language density. It is well known to biologists that species density per unit area is
highest in equatorial regions and declines towards the poles – a pattern known as
Rapaport’s rule – and languages obey it too. It is possible that the one causes the other,
that in some way higher biodiversity is capable of supporting greater cultural diversity,
but the explanation seems to be that both biological and cultural diversity depend on
the same environmental factors such as temperature and rainfall (Nettle 1999,
Moore et al. 2002, Sutherland 2003).
San hunters, speakers of a Khoi-San language, Namibia.
Unlike other language families, Khoi-San languages use clicks
made by the tongue and sharp intakes of air. The Hadza people
(front cover), 2,500 km away in Tanzania, also speak a click
language (see gure 9). Genetic evidence suggests that the most
recent common ancestor of these two peoples lived as long as
50-70,000 years ago, around the time that modern humans left
Africa. It is possible that the very rst languages ever spoken
were click languages, and that clicks evolved before vocal words
as a means of communicating without scaring animals when
hunting (Pennisi 2004).
© Martin Harvey / WWF-Canon
DIVERSITY The extinction crises facing
both species and languages as
consequences of similar processes
Until now, the story has been about evolution and diversication. There is another side
to the story, decline and extinction. Most species that ever existed have gone extinct.
They are the inner branches and twigs of the tree that stopped growing (or evolved into
another species) before reaching the outermost edge. Over and above the background
extinction rate, there have been at least ve biological mass extinction events since the
Cambrian explosion, in which global species diversity was suddenly reduced.
The third event was the greatest, 245 million years ago, in which 96% of species went
extinct, and the fth, 65 million years ago, marked the demise of the dinosaurs.
Following each mass extinction event, however, biodiversity recovered to or exceeded
its previous high level. The present rate of species loss may be in the region of
100-200 times higher than the background rate found in the fossil record (Groom-
bridge and Jenkins 2002), which puts us in the midst of a sixth mass extinction.
But this extinction event is cultural as well as biological (Nettle and Romaine 2000).
According to Ethnologue (Lewis et al. 2013), a periodic publication dating back to the
1950s which compiles data on the world’s languages and speaker numbers, half of
the world’s population speaks one of only 24 languages, the top ten being Mandarin
Chinese, Spanish, English, Hindi, Portuguese, Bengali, Russian, Japanese, Javanese
and German.10 These two dozen languages have speakers numbering in tens or
hundreds of millions. The other half of the world’s population speak the remaining
7,000 languages (see Figure 3).
The Ethnologue data describe an enormously skewed distribution of speakers among
the world’s languages. Figure 4 shows that the frequency of languages of dierent
sizes forms a normal, bell-shaped curve, but on a log scale – each category along the
horizontal axis of the graph is ten times the size of the previous one. Around half of the
world’s languages has fewer than 10,000 speakers, and the other half has more than
10,000. But 95% of the world’s population are found in the three size classes at the
right-hand end of the bell curve, they speak languages spoken by millions, tens of
millions or hundreds of millions of people. Forty percent of us occupy the tiny group
of languages with 100 million-plus speakers. At the other end of the distribution,
just over one percent of the world’s population are responsible for maintaining over
5,000 languages, those with fewer than 100,000 speakers. Astonishingly, only about
0.1% of the world population or about 8 million people, equivalent to a city about the
size of London, are responsible for keeping one half, or about 3,500, of the world’s
Number of languages
Number of speakers
LESS THAN TEN
TENS OF THOUSANDS
HUNDREDS OF THOUSANDS
TENS OF MILLIONS
HUNDREDS OF MILLIONS
3500 languages, 0.1% of world population 400 languages, 95% of world population
40% of population
Sizes of languages
Number of languages by size
class (after Harmon 1995,
2002; data source:
Lewis et al. 2013).
EGYPTIAN SPOKEN ARABIC
REST OF WORLD
More than half the world’s
population speaks one of
just two dozen languages
(source: Lewis et al. 2013).
A comparison with the distribution of language sizes two decades ago (Harmon 1995)
reveals that, while the world’s population has grown by about 25%, the number of
million-plus languages has expanded but small languages have dwindled away.
The distribution is gradually shifting to the right like a wave and becoming even more
skewed. The only group on the left of the graph that has grown is the zero (extinct)
class. Some linguists predict that 90% of the world’s languages will die out this century
(Nettle 1999, Nettle & Romaine 2000). Why is this happening?
Languages can go extinct either because the entire population of speakers dies out or,
more usually, because the speakers shift to a dierent language and, typically within
a few generations, forget their mother tongue. This can happen for social or economic
reasons, such as commerce or migration, or through a deliberate policy of linguistic
unication by a dominant group (see box: Linguistic Ecology). The globalization of
trade and media, and technological progress in transport and communication, have
accelerated the process of language shift, as have nationalization policies that favour
a small number of languages, increasing the pressure on languages with thousands or
fewer speakers, and boosting the dominance of those with millions. As language is the
primary medium of cultural transmission, linguistic diversity and cultural diversity are
being diminished simultaneously.
Most of the languages threatened with extinction are evolutionarily quite distinct from
the few dominant world languages, and so they also represent very dierent cultures.
Nearly all are spoken by indigenous people, some still living in traditional ways on
their ancestral lands, although these are becoming rare. Along with the languages,
the traditional knowledge of these indigenous cultures is being forgotten. The names,
uses and preparation of medicinal and food species, both plant and animal,
and traditional methods of farming, shing, hunting and natural resource management
are disappearing, not to mention the vast array of spiritual and religious beliefs and
practices that are often associated with traditional land use and resource management,
which are as diverse and numerous as the languages themselves. This vast store of
knowledge that has evolved and accumulated over tens of thousands of years could be
lost in the course of just two centuries, the 20th and the 21st. While linguists have made
great eorts to document, record and archive as many of the endangered languages as
possible, and ethnobiologists have attempted to record the traditional uses of plants
and animals by indigenous peoples, the most important conservation takes place on
the ground, as part of a living culture.
Conserving linguistic and cultural diversity presents a quite dierent ethical problem
compared with the conservation of biodiversity. There are very strong utilitarian and
economic arguments for protecting species and maintaining natural ecosystems,
but there is also a moral argument that no species should be extirpated for human
purposes. Cultures and languages on the other hand can only be maintained by people
who choose to, usually but not necessarily the ethnic group with which the culture
evolved, nobody should be forced to speak a language or practise a culture if they do
not want to. Most indigenous peoples, of course, do want to keep their language and
culture alive, but they may not have the opportunities or means or numbers to
LANGUAGES GO EXTINCT
EITHER BECAUSE THE
OF SPEAKERS DIES OUT
OR, MORE USUALLY,
BECAUSE THE SPEAKERS
SHIFT TO A DIFFERENT
TYPICALLY WITHIN A
FORGET THEIR MOTHER
Himba woman, a speaker of a Niger-Congo language. e Himba are semi-nomadic pastoralists who, unlike many indigenous groups in Africa,
have managed to maintain much of their traditional lifestyle. Kunene Region, Namibia.
© Martin Harvey / WWF-Canon
When linguists discuss the ecology of a region or a country they are not thinking of the
relationships between its species and their environment. They are talking about the
languages spoken in an area and the dynamics of the interactions between them and
the social and political context in which they exist. One of the dominant forces in lin-
guistic ecology is language shift. Language shift occurs when a population of speakers
adopts a new language at the expense of their mother tongue, generally over the course
of a few generations, and is the biggest driver of language extinction.
A well-documented, on-going example of the process of language shift comes from
Britain, where Scottish Gaelic11 has been losing speakers to English over the last 200
years (MacAulay 1992).12 In the mid-18th century the population of Scotland was
around 1.25 million, consisting of about 300,000 Gaelic speakers, concentrated in the
Highlands and Islands, and nearly one million English speakers, concentrated in the
Lowlands. The Highland clearances, a programme of removing small-scale farmers
from their land to make room for large-scale sheep farmers, and the consequent
migration of Gaelic speakers to the Lowlands or away from Scotland altogether led to
a steady decline in their number. By the end of the 19th century, monolingual Gaelic
speakers had mostly disappeared, and nearly all the remaining Gaelic speakers were
bilingual. Today the population of Scotland is around ve million, with about 58,000
Gaelic speakers, just above one percent of the total (note that Scottish Gaelic is still
above the world median language size, and therefore one of the worlds’ larger lan-
guages). The Scottish government has made eorts to promote primary education in
Gaelic and, although the number of Gaelic speakers continued to decline between
2001 and 2011, the number of speakers aged under 20 remained stable.
1750 1800 1850 1900 1950 2000
Speakers of English only
Gaelic speakers (mono and bilingual)
Of the other Celtic languages, Irish, the closely-related sibling of Scottish Gaelic, is
declining alongside it, and Breton (spoken mainly in Brittany, France) is declining
faster. The last mother tongue Cornish13 speaker died in 1777 and the last Manx14
speaker in 1974, although attempts are being made to keep them alive as second
Welsh is the only Celtic language with a strong speaker base owing to decades of
support from the educational system and government policy. Celtic languages have
been struggling along beside far larger, socially and politically dominant languages,
English and French, for more than a thousand years. The British Isles had an entirely
Celtic-speaking population up until the time of the Roman invasion, and remained
predominantly Celtic-speaking until the arrival of the Anglo-Saxons in the 5th century.
Gradually the Celtic languages were pushed to the western fringes where they survive
today. Language shift, to be very clear, does not mean that one population replaces
another, but that one language is displaced by another within the same population.
The peoples who spoke Celtic languages are still there, genetically the population is
still largely Celtic, even in England.
Nenets reindeer herdswoman, a speaker of a Uralic language,
eating reindeer meat, Kánin Peninsula, Russia, Arctic.
© Staffan Widstrand / WWF
LANGUAGES Global similarities and regional
dierences in the state of
biological and linguistic diversity
Because species and languages are alike in terms of their evolution, diversity, and dis-
tribution around the world, it is appropriate and feasible to assess their current status
in similar ways, and compare the two. We have adapted and applied two methods
developed for assessing the state of biodiversity to measure the state of linguistic diver-
sity. The rst is the IUCN Red List system which is used to assess the extinction risk to
species (IUCN 2013); the second is the WWF/ZSL Living Planet Index which measures
the rate at which biodiversity is declining (Loh et al. 2005, Collen et al. 2009).
Threat Status of Species – Red Listing
The IUCN Red List is a system used by biologists to assess the conservation status of
plant and animal species. It is based on a set of categories for ranking species according
to their risk of extinction. There are seven categories ranging from Least Concern to Ex-
tinct. There is an eighth category for species which have been evaluated but for which
there are insucient data to assess their status. Those species which are categorised as
vulnerable, endangered or critically endangered are considered to be threatened.
EXTINCT IN THE WILD (EW)
CRITICALLY ENDANGERED (CR)
NEAR THREATENED (NT)
LEAST CONCERN (LC)
DATA DEFICIENT (DD)
NOT EVALUATED (NE)
IUCN Red List categories
Only a fraction of all species have been evaluated, but a few taxa (species groups)
have been completely evaluated and some have had a random sample of 1500 species
evaluated. Among the vertebrates, all mammals (5,501 species known and described to
date), birds (10,064 species) and amphibians (6,771 species) and a random sample of
1,500 species of all reptiles (approximately 9,000 species) have been recently assessed.
More than 10,000 sh species have been assessed, but as the sample is not random,
and the total number of sh species is very large but uncertain (about 32,000 known
to date), no rm conclusions can be reached about the status of sh as a group. These
assessments are used here to compare the threat status of vertebrate species with the
status of languages, based on a random sample of 1500 languages. For the purposes
of these comparisons, we combine the category extinct in the wild (EW) with extinct
(EX), and the category near threatened (NT) with least concern (LC).
The criteria used to categorise the conservation status of a species into one of the Red
List categories include a species’ population size, its rate of reduction (if in decline), its
range size and rate of decline or fragmentation, existing and future threats, or a com-
bination of these. It is possible to apply some of these criteria to languages and assess
their threat status according to either the number of mother-tongue speakers, their
rate of decline, or a combination of the two. Range size is harder to apply to languages
and therefore was ignored in this analysis, as was existing or projected threat. Because
biologists use a wider range of criteria to assess species than has been applied here to
languages, the threat status of languages should be considered more conservative.
Threat Status of Languages – UNESCO and Ethnologue
Linguists consider a language to be endangered if it is not being transmitted success-
fully from one generation of speakers to the next. This is very good reasoning, but it
means that the criteria used by linguists to assess the threat status of a language are
quite dierent to the IUCN criteria used by biologists. Ultimately the two sets of crite-
ria, linguistic and biological, are designed to assess extinction risk. Table 3 compares
the Red List criteria we have applied to a random sample of 1,500 languages with the
criteria used in two systems designed to assess threatened languages, UNESCO’s Lan-
guage Vitality and Endangerment system (UNESCO 2010) and Ethnologue’s Expanded
Graded Intergenerational Disruption Scale or EGIDS system (Lewis and Simons 2010)
The systems are not correlated: critically endangered in the Red List system does not
necessarily correspond to critically endangered in the UNESCO system for example;
the only category that means the same in all three systems is extinct.
Red List (as applied here) UNESCO Ethnologue (EGIDS)
No speakers remain.
No one can speak the language.
The language is no longer used.
Extinct in the Wild (EW):
The language serves as a reminder
of ethnic identity but no proﬁcient
Critically Endangered (CR):
Either the number of speakers is
observed or projected to decline by
80% or more in three generations
(75 years); or speakers number less
than 250 and declining by 25% or
more in one generation (25 years);
or speakers number less than 50.
Youngest speakers are great-
grandparents; language not used
on a regular basis; language only
Only spoken by great-grandparent’s
generation who have little opportunity
to use the language.
Either no. speakers observed or
projected to decline by 50% or more
in three generations (75 years);
or no. speakers less than 2,500
and declining by 20% or more in
two generations (50 years); or
no. speakers less than 250.
Language spoken only by grand-
parents’ and older generations;
parents understand but do use it
to speak to their children or
Only speakers are grandparents’
Youngest speakers are parents’
generation; children are not using
the language at home.
Parents’ generation use the language
among themselves but it is not being
transmitted to their children.
Either no. speakers observed or
projected to decline by 30% or more
in three generations (75 years); or
speakers number less than 10,000
and declining by 10% or more in
three generations (75 years); or
speakers number less than 1,000.
Most children speak their parental
language as their mother ﬁrst
language, but usage is restricted
to the home or particular social
The language is used by all genera-
tions, but it is losing users.
Near Threatened (NT):
The language does not meet the
criteria for CR, EN, or VU but is
likely to do so in the near future
(this category has not been used
in this assessment).
Stable yet Threatened:
The language is spoken by all
generations in most contexts, but
multilingualism is common and a
more dominant language is taking
over in some contexts.
Least Concern (LC):
The language does not fall into any of
the categories above; speakers are
widespread and abundant.
The language is spoken by all
transmission is uninterrupted.
The language is used by all
generations, and the situation is
Ethnologue further denes a number of higher categories for languages in vigorous
use: namely where standardized literature is in use but not widespread (Developing);
standardization and literature are in widespread use in education (Educational); the
language is used at work and in mass media but without ocial status as a national or
regional language (Wider communication); used in education, work, mass media and
government at provincial or national level (Provincial, National); used internationally
for trade, knowledge exchange or policy (International).
The UNESCO and Ethnologue EGIDS systems use inter-generational transmission as
the principal criterion in assessing a language’s vitality, dened according to the num-
ber of generations that speak the language: great-grandparents only, grandparents and
older, parents and older, or all including children. While there is an undeniable logic to
categories under three
systems of assessing the
status of languages
these systems, there are some good reasons for using the IUCN Red List system deve-
loped by biologists to assess the status of a language. Firstly, if children are no longer
speaking their parental language, unless there is great eort to revitalize the mother
tongue, it is inevitable that the language will move up through the categories towards
extinction. However, if a language that is close to extinction were to undergo a massive
revitalization eort, it would not move back down through the categories as rst the
grandparents, then parents and nally children learn to speak the language once again.
The linguistic categories assume there is one way trac up the ladder to extinction. But
it should be possible to track a reversal in the fortunes of a language dropping back down
the categories, which is the case if the Red List criteria are applied. Secondly, the status
of a language may change from location to location, or even from family to family, as
children could be speaking their mother tongue in some places, while only parents or
grandparents use the language in others. The Red List criteria are not concerned with the
age of the speakers, only the total numbers. Of course, the end result of a breakdown in
inter-generational transmission will be a decline in speaker numbers, so the Red List cri-
teria are focusing on the ultimate eect rather than the direct causes of endangerment.
The linguistic criteria recognize that a language may be safe or vigorous even if it is
only spoken by a very small population, as long as inter-generational transmission is
uninterrupted. The biological criteria conversely consider a language to be threatened
simply if the number of speakers is below a critical threshold (1,000 for vulnerable,
250 for endangered, 50 for critically endangered), even if there is no decline through
the generations. This is justiable as it is precisely when the mass of speakers is small
that a language could be threatened by a shift away from the mother tongue towards a
more dominant language by means of unforeseen events extraneous to the process of
intergenerational language transmission.
Comparison of Conservation Status of Languages and Species
Most importantly for our present purposes, applying the IUCN Red List criteria to
languages allows us to assess their threat status on the same basis as species, and make
comparisons on a quantied basis. This has been done previously by the ecologist
William Sutherland (2003), who used a limited set of the Red List criteria to com-
pare languages with birds and mammals. Sutherland found that a higher percentage
of languages was either threatened or recently extinct (32%) than either birds (13%)
or mammals (28%). Here we compare the status of languages with that of mammals,
birds, reptiles and amphibians, and compare the status of languages between
regions of the world and between dierent language families, using selected
Red List criteria. The data on numbers of speakers of languages come from editions
of Ethnologue dating from 1951 to 2009, although most of the data come from the
1990s and 2000s (see Table 4).
Period Data points
Number of data points
on speaker numbers
Past data are sometimes unreliable, especially in the earlier decades, and therefore the
rst Red List criterion – rate of decline observed or projected over three generations –
has not been used in this analysis. Trends in speaker numbers have only been used in
combination with total number of speakers (the second criterion). Therefore the
assessment of languages is very much more conservative than that of species groups.
The results are shown in Figure 7.
The analysis indicates that at least a quarter of the world’s languages are threatened
with extinction (CR, EN or VU), assuming that no data decient (DD) language is
threatened, compared with at least 21% of mammals, 13% of birds, 15% of reptiles and
30% of amphibians, the most threatened class of vertebrate. Furthermore, about 6% of
languages have been reported as recently extinct, as opposed to about 1% of vertebrate
species. If sucient data on all the criteria used to evaluate animals were available to
assess languages, then the status of languages could be worse than it appears here.
Ethnologue reports gures for the numbers of languages in each EGIDS category
except for Extinct. They are Dormant 2.9%, Nearly Extinct 6.0%, Moribund 4.1%,
Shifting 6.5%, Threatened 14.8%, Vigorous or better 65.7% (Lewis et al. 2013).
If the EGIDS categories were translated into Red List categories as in Table 3,
the percentages would be quite similar to those given in Figure 7.
Red List conservation
status of languages and
four vertebrate classes
Size of each pie is
proportional to the
number of languages or
species in each group
Mammal, bird and
amphibian data from
IUCN (2013), reptile data
from Bohm et al. (2013).
Extinct (since 1970)
Least Concern/Near Threatened
Red List Status
1% 2% 4%
1% 4% 8%
Threat Status of Language Families and Regions
Just as the world’s population is not evenly distributed among the world’s languages,
with half the world speaking one of just 24 languages, so the world’s languages are not
evenly distributed among language families. Figures 8a and 8b show the dominance
of a few major language families such as Afro-Asiatic, Austronesian, Indo-European,
Niger-Congo and Sino-Tibetan.
language families by
number of languages
(source: Lewis et al. 2013)
144 123 102 313
The largest language
families by number of
(source: Lewis et al. 2013)
Status of Language Families
Some of the larger language families from dierent regions of the world have been
assessed to compare their conservation status, in exactly the same way as all languages
were assessed as a whole. Figure 9 shows the percentage of languages in each Red List
category. The status of languages in each family can be compared by looking at the
percentage of languages in the extinct (EX) and threatened categories (CR, EN, VU).
Note that this assumes that all data decient (DD) languages are not threatened, so it is
a conservative estimate of threat status.
Extinct (since 1970)
Least Concern/Near Threatened
Language Families (source: Lewis et al. 2013)
Conservation status of language
(size of pie is proportional to number
It is clear that languages in the Australian family are the most severely endangered, with
94% of languages threatened with extinction or extinct (since 1970), followed by some
of the American language families such as Na-Dene (73%), Tupian (69%) and Arawakan
(65%). The least endangered families are Niger-Congo, with only 6% of languages threat-
ened or extinct, Nilo-Saharan (8%) and Indo-European (9%). Other families, such as
Altaic (29%), Austronesian (29%) and Afro-Asiatic (24%) show moderate levels of threat.
1% 3% 7%
4% 2% 4%
3% 3% 8%
6% 6% 12%
10% 3% 3% 3%
2% 2% 5%
Status of Regions
A clear pattern emerges if the data are analysed by region rather than family. Figure 10
show the percentage of languages in each Red List category. The status of regions
can be compared by adding up the percentage of languages in the extinct (EX) and
threatened (CR, EN, VU) categories. As with families, this assumes that all data
decient (DD) languages are not threatened, so it is a conservative estimate of threat
status of each region.
Extinct (since 1970)
Least Concern/Near Threatened
Conservation status of
languages by region
(size of pie is proportional to the
number of languages in each region)
3% 2% 2% 4%
Languages of the Pacic and the Americas are the most severely endangered,
both regions with about 60% of their languages threatened with extinction or extinct
(since 1970). The regions with the lowest level of extinction risk are Africa (11%)
and Asia (20%). If Australia is separated from the rest of the Pacic region, it is once
more apparent that Australia’s languages are the most severely endangered in the
Paciﬁc (incl. Australia and PNG)
4% 2% 6%
Trends in Languages and Species
Another way that biologists assess the state of biodiversity is to use indices based on
average trends in the populations of a selection of species, such as the WWF/ZSL
Living Planet Index (LPI). Species population indices are essentially like stock market
indices such as the Dow Jones or Financial Times Stock Exchange (FTSE) which
track trends in market capitalization of a number of companies. The LPI is based on
time-series data for approximately 9,000 vertebrate species populations (of about
2,600 dierent species) from around the world. The index has been published
biannually by WWF and ZSL since 1998 (WWF 2012).
In previous work, the authors adapted the LPI method to create an index called the
Index of Linguistic Diversity (ILD) (Harmon and Loh 2010) which can be compared
with trends in biodiversity as measured by the LPI. The ILD uses trends in the num-
bers of speakers across a sample of languages to calculate average trends. The same
sample of 1,500 languages used in the Red List analysis of languages was also used to
calculate the ILD. Data on numbers of mother tongue speakers for each language going
back to 1900 were extracted from editions of Ethnologue. After removing all languages
with data from only a single point in time (412 languages), and then ltering the data
to remove anomalous data points or time series,16 the remaining dataset contained
time series for 985 languages. The ILD calculates the average trend of those languages
in the sample. The ILD results for all 985 languages is a measure of trends in linguistic
diversity for all languages in the world, and may be compared with the global LPI to
see relative trends in linguistic diversity and biodiversity. To facilitate regional compa-
risons between the two indices, we also calculated the ILD by biogeographic realm to
match up with the way the LPI is calculated regionally.
The biogeographic realms used in the analysis are the Afrotropical, the Indo-Pacic,
the Nearctic, the Neotropical and the Palearctic realms. These are regions of the world
dened according to the shared evolutionary history of their biota. It is a useful way to
compare trends with languages, as language families tend to conform approximately to
the same biogeographic patterns. Table 5 below shows which families belong to which
Biogeographic Realm Location Language families analyzed in this report
Afro-tropical Sub-Saharan Africa Niger-Congo, Nilo-Saharan
Indo-Paciﬁc South and Southeast Asia including southern
China, Australasia and Oceania
Australian, Austro-Asiatic, Austronesian,
Nearctic North America and a part of northern Mexico Na-Dene
Neotropical Latin America and the Caribbean Arawakan, Mayan, Oto-Manguean, Tupi
Palearctic Eurasia, northern Africa and the Middle East,
Central Asia, northern China and Japan
Afro-Asiatic, Altaic, Indo-European, Sino-
Distribution of some
The major dierence between the LPI and ILD methodology is that the ILD has been
corrected for overall human population growth. Within the period covered by the
index, the human population has more than doubled, whereas there is no comparable
overall growth in global wildlife populations. Therefore all of the ILD graphs presented
here have been corrected for overall population growth. The biogeographic realm ILDs
have been corrected for human population growth within the realm boundaries. The
ILD therefore is not an index of population trends in quite the same way as the LPI;
what the ILD measures is trends in the fraction of the total population belonging to
each language.17 To use another economic analogy, it is like an index of average market
share of languages. If the average market share declines, it means that a few languages
are increasing their market share at the expense of a greater number of others. This is
exactly what we would expect to see if language shift is taking place: as speakers shift
away from many small languages to fewer larger languages, then the average market
share index falls.
The ILDs and LPIs for each biogeographic realm are shown in gure 12, plotted on the
same axes for comparison, as are the global indices. Because the Indo-Pacic realm
includes two islands which are particularly important in terms of linguistic diversity,
Australia and New Guinea, additional ILDs for the two are shown separately in gure
11. Australia shows the fastest decline in linguistic diversity of any country, with a fall
of about 85% in its ILD from 1970 to 2009. The ILD for New Guinea, the number one
hotspot for linguistic diversity, which includes the Indonesian half of the island plus
the half that is Papua New Guinea (PNG), shows a decline of about 40% between
1970 and 2005. This is a faster decline than the global average ILD, and reects the
Red List status of the island which shows that over 50% of New Guinea’s 1000 or
more languages are threatened.
1970 1975 1980 1985 1990 1995 2000 2005 2010
New Guinea (incl. PNG and West Papua)
Index of Linguistic
Diversity. Australia and
Global Index of
(ILD) and Living Planet
Trends in the LPI and ILD
Globally, the indices of both species populations (LPI) and speakers of languages (ILD)
are declining at similar rates, about 30% in 40 years. The most rapid declines in
species since 1970 have occurred in the Afrotropics (about 40%), Indo-Pacic (about
65%) and Neotropics (about 50%), whereas the Nearctic and Palearctic have shown
little overall change. For languages, the most rapid declines since 1970 have taken
01970 1975 1980 1985 1990 1995 2000 2005 2010
01970 1975 1980 1985 1990 1995 2000 2005 2010
01970 1975 1980 1985 1990 1995 2000 2005 2010
place in the Nearctic and Neotropical realms (both about 75%), whereas the rate of
decline in the Afrotropical (about 20%), Indo-Pacic (about 30%) and Palearctic
(about 30%) realms has been slower. In summary, biodiversity has declined rapidly in
the tropics, but remained steady in temperate realms; linguistic diversity on the other
hand has declined rapidly in the new world, but more slowly in the old world.
01970 1975 1980 1985 1990 1995 2000 2005 2010
01970 1975 1980 1985 1990 1995 2000 2005 2010
01970 1975 1980 1985 1990 1995 2000 2005 2010
It is striking that biodiversity and cultural diversity in general, and species and
languages in particular, show extraordinary parallels both in terms of their evolution
and the threats they face. Both species and languages have evolutionary histories which
can be traced back in time to earlier, ancestral species and languages; both languages
and species can be classied in such a way as to show the phylogenetic relationships
between those related by descent from a common ancestor. Species may be dened by
the ability to interbreed; language may be dened by mutual intelligibility.
By these denitions, subspecies are analogous to dialects. The process of formation
of new species, speciation, has its linguistic equivalent. It may even be argued that the
evolutionary mechanisms that give rise to both species diversity and linguistic diversity
are similar. Biological evolution and, it can be argued, cultural evolution are the result
of the action of replication, variation and selection working on hereditary material.
As well as the genetic relationships between languages, linguists talk about the
ecology of a region or country, which has nothing to do with ora or fauna.
The global distribution of languages and species show remarkable similarity,
with diversity highest in the tropics and declining toward the poles.
Two results are immediately apparent when comparing the status and trends in
biodiversity and linguistic diversity. Firstly, at the global level, the trends are very
similar, both the LPI (species) and ILD (languages) declined by about 30% since 1970,
which suggests that biodiversity and linguistic diversity are being lost at similar rates.
This supports the conclusion of the Red List analysis comparing the conservation
status of languages and species: globally, linguistic diversity is at least as threatened
The second result is that, while both biodiversity and linguistic diversity are
threatened globally, they are declining at dierent rates in dierent regions of the
world. By far the most rapid losses in linguistic diversity have occurred in the Americas
where, according to the Red List analysis, 60% of languages are threatened or have
gone extinct since 1970. The ILD plummeted by over 75% between 1970 and 2009
in both the Nearctic and the Neotropical biogeographic realms. The LPI, however,
shows that while species populations have fallen in the Neotropics (although with high
uncertainty limits), they were almost completely at in the Nearctic. The LPI fell by
more than 60% in the Indo-Pacic, whereas the ILD declined by about 30%, a similar
rate to the global average. Of course this masks the catastrophic decline of more than
80% in the ILD of Australia (and more than 40% in New Guinea). The ILDs for the
Afrotropical and Palearctic realms both show declines of around 20-30%.
The dierence in regional trends between the LPI and the ILD can be explained
by the dierent direct pressures faced by biodiversity and linguistic diversity.
Biodiversity decline is the usually the result of one of ve main direct threats or
pressures: habitat loss and destruction, direct over-exploitation of species from
hunting and shing, competition or predation by invasive alien species, climate
change, or pollution. Habitat loss and over-exploitation of species remain the greatest
threats for most of the world’s biodiversity, and over the last 40 years the strongest
pressure has been felt in the tropics, especially in Asia.
THE DIFFERENCE IN
BETWEEN THE SPECIES
AND LANGUAGES CAN
BE EXPLAINED BY THE
PRESSURES FACED BY
In Europe and North America most of the biodiversity loss from habitat destruction
occurred before 1970 and so does not register on the LPI. However, the footprint
of natural resource consumption by the developed world is felt increasingly in the
developing tropics, and all the more so as China’s demand for natural resources grows;
so it is the growing consumption by rich counties as well as population growth in poor
countries that are driving the loss of tropical biodiversity in Africa, Asia and
The decline in linguistic diversity is normally a result of the process of language shift
away from small indigenous languages toward larger, national or regional languages.
Language shift is driven by a number of social, political and economic factors including
migration, urbanization, national unication, colonization, and the globalization of
trade and communications. Migrant communities often undergo a process of language
shift, whether moving from one country to another, or from a rural to an urban area
within the same country. Governments in many developed and developing countries
actively promote a single national language at the expense of other, usually minority,
languages for political reasons. This has been the case with Mandarin in China,
French in France and Amharic in Ethiopia for example. Migration, urbanization and
political nationalization have been the primary drivers in Africa, Asia and Europe,
where language shift has tended to occur between languages within the region.
In the Americas and the Pacic, especially Australia, the primary driver has also been
migration, but the migrants, mainly European, vastly outnumbered the indigenous
populations, and so it was the migrants’ languages, primarily English, Spanish and
Portuguese, that became politically and economically dominant. It is in these regions
where indigenous languages are most highly threatened.
Australia and the island of New Guinea deserve particularly close attention: Australia
because its indigenous languages are the most highly threatened in the world, and New
Guinea because it is the most linguistically diverse place on Earth. Most of the 1,000
or so languages of New Guinea are threatened, but their decline is not as rapid as in
Australia where more than 90% are threatened with extinction. The dierence between
the two islands is of course due to the fact that the vast majority of the Australian
population is of European descent, whereas the population of New Guinea is largely
indigenous. Australian languages are spoken by minority indigenous communities,
and among these communities English is taking over, or has taken over, as the rst
language. In New Guinea indigenous languages are faring better, although the English-
derived lingua-franca Tok Pisin is gaining ground at their expense.
MIGRATION, URBANIZATION AND POLITICAL NATIONALIZATION HAVE BEEN THE
PRIMARY DRIVERS OF LANGUAGE LOSS IN AFRICA, ASIA AND EUROPE. IN THE
AMERICAS AND AUSTRALIA, THE PRIMARY DRIVER HAS ALSO BEEN MIGRATION,
BUT THERE THE MIGRANTS, MAINLY EUROPEAN, VASTLY OUTNUMBERED THE
FOOTPRINT OF RICH
COUNTIES AS WELL AS
IN POOR COUNTRIES ARE
DRIVING THE LOSS OF
IN AFRICA, ASIA AND
A Bajau (“sea gypsy”) woman and children, speakers of an
Austronesian language, Kusungan Island, Sabah, Malaysia.
Bajau people originated from the Philippines and traditionally
lived on boats, making their living from the sea, but most are
© WWF-Malaysia / Mazidi Abd Ghani
A future for biocultural diversity?
Ultimately, both linguistic diversity and biodiversity are diminishing as a result of
human population growth, increasing consumption, and globalization which erodes
dierences between one part of the world and another. At the regional level, these
fundamental drivers of diversity loss are manifest in dierent ways. For biodiversity,
the biggest threat in modern times has been and still is habitat destruction, followed by
over-exploitation (shing and hunting) and invasive species. Since 1970, habitat loss
has been most rapid in the developing world, particularly in Asia, whereas habitat
loss in Europe and North America has slowed down and levelled o. Consequently the
most rapid decline in biodiversity is now happening in the tropics, the part of the world
with the greatest diversity.
For languages and culture, ‘habitat’ means the human population, which has doubled
since 1970, so habitat loss is not the problem. Nor does direct ‘consumption’ threaten
culture (there is no equivalent of cultural over-consumption). It is the cultural analo-
gue to alien invasive species – language shift – that is the greatest threat to linguistic
and cultural diversity. It is not that one human population replaces another populati-
on, as is the case with invasive species, it is that one language displaces another langu-
age within the same population. When an indigenous language goes extinct, often the
indigenous culture follows. This process has been happening for the last two hundred
years or more in the linguistic ecology of the Americas, Australia and parts of the Pa-
cic, where indigenous languages have been severely threatened by the dominance of
European languages, particularly English, Spanish and Portuguese. In Africa, Asia and
Europe, where the main drivers of language shift have been migration, urbanization
and political unication policies, language shift has tended to occur between languages
of those regions, and diversity is being lost, but not as rapidly.
Why do we need so much diversity? Would it not be better for the sake of world peace
and the global economy, it is sometimes argued, if we spoke fewer languages in the
world? Are languages or cultural diversity really worth conserving as much as species
or biological diversity? The logical conclusion of this type of argument is that, ideally,
we should speak just one world language. But then we would all become more similar,
and the dierences between one part of the world and another, or between one culture
and another, would rapidly erode away. In the end, we would speak the same language,
wear the same clothes, eat the same food, listen to the same music, consume the same
brands and hold the same beliefs. One city would look much the same as another.
The world would become homogenized. This counter-argument may sound absurd,
but it is already happening: the world is already losing its extraordinary biocultural
diversity, as the ndings of this report demonstrate. No doubt the global economy
would continue to grow just as well, or even better, with just a few world languages
and cultures. It is even possible that global ecosystems could continue to provide basic
life support functions – although probably not as well – with less biodiversity, and
humanity would still survive. But this is not just a question of survival, or even global
economic productivity. A diverse world is a culturally and naturally richer world.
With less diversity, humanity is poorer. It is a question of the kind of world we want
to live in.
DIMINISHING AS A
RESULT OF HUMAN
BETWEEN ONE PART
OF THE WORLD AND
The science of biocultural diversity is in its infancy, and more research is needed to
examine and understand the processes and mechanisms that underpin and unite
biological and cultural evolution and ecology. Most importantly, we need a better
understanding of how to slow down and reverse the loss of diversity. While the outlook
is not bright for many of the world’s smaller languages, especially those no longer
being learned by children, there is plenty of scope to improve, develop and promote
biocultural conservation. There is an opportunity for biodiversity conservation and
the conservation of indigenous languages and cultures to go hand in hand. Most of the
world’s linguistic diversity is found in areas of high species richness and endemism.
If biodiversity conservation organizations on the ground in areas of high biocultural
diversity were to invest resources in the conservation of indigenous languages and
traditional knowledge there would be a double pay-o. Field linguists working
on indigenous languages often lack the ecological knowledge needed in order to
understand and translate the vast lexicon of terms for species and natural phenomena.
Field biologists could benet from the immense wealth of traditional ecological
knowledge of indigenous people. Not only would biological and cultural diversity be
conserved together in the environment in which they both evolved, so protecting the
full range of living biocultural diversity, but also the traditional resource management
systems, a fundamental component of the cultural identities that are now in retreat,
could be applied to conserving the landscape, its component species and its languages.
But it is not only the rarest languages and species that we should conserve. Relatively
common languages, spoken by tens of thousands of people, and common species are
in decline too. Maintaining diversity is not just a question of protecting endangered
languages and species in remote hotspots of biocultural diversity such as the Amazon
or New Guinea, vitally important though that is, conservation is also a matter of
allowing diversity to thrive in those parts of the world where humans have already
had a profound impact on the biological and cultural landscape, in the more densely
populated parts of the planet. Recognizing and exploring the parallels between nature
and culture, and understanding the processes that underlie their evolution, ecology
and extinction, is a rst step towards ensuring that we can continue to inhabit a world
of incredible diversity.
Mongolian herder, speaker of an Altaic language, Baga Lake,
Khar Us Nuur National Park, Mongolia.
© Frans Schepers / WWF-Netherlands
1 It is possible for a person to speak more than one language, of course, whereas an individual
animal or plant cannot belong to more than one species. In this report, we refer to mother-
tongue speakers of a language only. This means one’s native language, or the language one most
strongly identies with as a native speaker, which is usually, although not necessarily, the
language one learned rst. When we report the numbers of speakers of a language, we mean
mother-tongue speakers, so each speaker is only counted once.
2 There are actually quite a number of dierent denitions of “species” and “languages,” and
multiple processes of speciation and language genesis. For details, see Harmon 2002.
3 This description of the evolution of biocultural diversity builds on JL’s contribution “The third
owering of the Tree of Life” to Jorgen Randers, 2052 - A Global Forecast for the Next Forty
4 Ours is not the only species to have culture – many species of songbird, for example, show
regional and local variations in their repertoire of songs which are not inherited genetically but
learned from other individuals of the same species, and some species of primate even have a
limited repertoire of calls with specic meanings such as “snake” or “leopard” – but we are the
only species to have language.
5 French belongs to the Italic branch, which is also part of the Indo-European family.
6 There is still controversy surrounding the exact dates and routes of the human diaspora out
of Africa. The description here is based on Oppenheimer (2004).
7 It is likely that proto-language, consisting of sounds, gestures and expressions, had begun
evolving long before that time.
8 Strictly speaking, this is West Frisian, as there are two other Frisian languages spoken in
9 Basque is a language isolate (in a family of its own) in northern Spain and southwest France,
while Navajo is a language in the indigenous Na-Dene family spoken in the southwest
10 Arabic is not included in the top ten as it is classied as many dierent languages, such as
Algerian spoken Arabic, Egyptian spoken Arabic, etc. If all speakers of Arabic languages were
counted together, Arabic would appear in the top ten.
11 Scottish Gaelic, a Celtic language related to Irish, should not be confused with Scots, a Germanic
language related to English.
Native Scots Gaelic speakers and bilingual English-Gaelic speakers are counted together in
13 Cornish, a Celtic language related to Welsh and Breton, spoken in Cornwall.
14 Manx, a Celtic language related to Irish and Scottish Gaelic, spoken on the Isle of Man.
15 A few of Australia’s languages are not in the Australian language family, hence the dierence
between the percentage of Australia’s languages that are extinct or threatened with extinction
and that for Australian languages.
16 Languages with 1,000 or more speakers which grew or declined at a rate greater than
10% per year.
17 For more detailed discussion of the ILD, see Harmon and Loh (2010).
Language or Species Group EX CR EN VU LC DD Total assessed
Languages (sample of 1500) 6% 7% 7% 11% 63% 5% 1.500
Mammals (all) 1% 4% 8% 9% 63% 15% 5.506
Birds (all) 1% 2% 4% 7% 85% 1% 10.065
Reptiles (sample of 1500) 0% 2% 6% 7% 64% 21% 1.500
Amphibians (all) 1% 8% 12% 10% 44% 25% 6.409
Language Family EX CR EN VU LC DD Total assessed
Afro-Asiatic 11% 5% 4% 4% 73% 3% 75
Altaic 0% 6% 12% 12% 65% 6% 17
Australian 33% 41% 18% 2% 2% 4% 51
Austro-Asiatic 0% 3% 8% 6% 81% 3% 36
Austronesian 2% 4% 7% 17% 67% 4% 245
Indo-European 2% 0% 2% 5% 79% 13% 104
Niger-Congo 0% 1% 1% 4% 89% 5% 296
Nilo-Saharan 3% 3% 0% 3% 82% 10% 39
Sino-Tibetan 0% 1% 3% 7% 77% 11% 87
Trans-New Guinean 2% 5% 17% 26% 45% 6% 119
Arawakan 30% 5% 10% 20% 35% 0% 20
Mayan 0% 7% 0% 13% 73% 7% 15
Na-Dene 27% 36% 9% 0% 18% 9% 11
Oto-Manguean 3% 0% 25% 8% 53% 11% 36
Tupi 25% 13% 13% 19% 25% 6% 16
Region EX CR EN VU LC DD Total assessed
Africa 3% 2% 2% 4% 87% 1% 402
Americas 17% 16% 12% 14% 38% 2% 255
Asia 2% 2% 6% 9% 77% 3% 464
Europe 11% 5% 0% 7% 70% 7% 44
Paciﬁc (incl. Australia and PNG) 7% 13% 15% 26% 38% 1% 282
Australia 32% 40% 17% 4% 6% 2% 53
New Guinea (incl. PNG and West Papua) 1% 6% 16% 29% 48% 0% 220
Biogeographic Realm Index Lower conﬁdence limit Upper conﬁdence limit
Global 28% 31% 26%
Afrotropical realm 22% 25% 19%
Indo-Paciﬁc realm 29% 32% 26%
Nearctic realm 74% 78% 71%
Neotropical realm 73% 79% 67%
Palearctic realm 28% 34% 21%
Australia 86% 89% 82%
New Guinea (1970-2005) 41% 46% 38%
Conservation status of
languages and species
Data for mammal, bird and
amphibian species from IUCN
2013, data for reptiles from
Bohm et al. 2013, percentage
in each Red List category.
Conservation status of
Percentage of languages in
each Red List category.
Conservation status of
languages by region
Percentage of languages in
each Red List category.
Index of Linguistic
Diversity global and by
Böhm, M., B. Collen, J.E.M. Baillie, et al. 2013. The conservation status of the world’s reptiles.
Biological Conservation, 157, 372–385
Cavalli-Sforza, L.L. 2000. Genes, People and Languages. Penguin, London.
Collen, B., J. Loh, S. Holbrook, L. McRae, R. Amin & J.E.M. Baillie. 2009. Monitoring Change in
Vertebrate Abundance. Conservation Biology, 23 (2), 317-327.
Darwin, C. 1874. The Descent of Man, and Selection in Relation to Sex. Second edition. John
Murray, London. Downloaded 12 February 2014 from van Wyhe, J. (ed.) 2002. The Complete
Work of Charles Darwin Online http://darwin-online.org.uk
Dawkins, R. 2006. The Selsh Gene. 30th Anniversary Edition. Oxford University Press, Oxford.
Dennett, D. 2002. The New Replicators. In: M. Pagel (ed.-in-chief) Encyclopedia of Evolution, 1,
E83-92. Oxford University Press, Oxford.
Gray, R.D. & Q.D. Atkinson. 2003. Language-tree divergence times support the Anatolian theory
of Indo-European origin. Nature, 426, 435-439.
Gray, R.D., Q.D. Atkinson & S.J. Greenhill. 2011. Language evolution and human history: what a
dierence a date makes. Phil. Trans. R. Soc. B, 366, 1090-1100.
Greenberg, J. 2000. The Eurasiatic Language Family: Indo-European and its Closest Relations.
Volume 1: Grammar. Stanford Univ. Press, Stanford, California.
Groombridge, B. & Jenkins, M.D. 2002. World Atlas of Biodiversity. Prepared by UNEP World
Conservation Monitoring Centre. Univ. California Press, Berkeley.
Harmon, D. 1995. The Status of the World’s Languages as Reported in Ethnologue. Southwest
Journal of Linguistics, 14 (1-2), 1-28.
Harmon, D. 1996. Losing Species, Losing Languages: Connections between Biological and Lin-
guistic Diversity. Southwest Journal of Linguistics, 15 (1-2), 89-108.
Harmon, D. 2002. In Light of Our Dierences: how diversity in nature and culture makes us
human. Smithsonian Institution Press, Washington and London.
Harmon, D. & J. Loh. 2010. The Index of Linguistic Diversity: A New Quantitative Measure of
Trends in the Status of the World’s Languages. Language Documentation and Conservation, 4,
97-151. Available online http://nrc.hawaii.edu/ldc/2010/
Hrozný, B. 1917. Die Sprache der Hethiter: ihr Bau und ihre Zugehörigkeit zum indogermani-
schen Sprachstamm: ein Entzierungsversuch (The Language of the Hittites: its Structure and
its Membership in the Indo-European Linguistic Family). J.C. Hinrichs, Leipzig, Germany.
Human, S. no date. Mapping the Genetic Relationships of the World’s Languages. http://www.
worldgeodatasets.com/language/human/ Downloaded 22 February 2012.
IUCN 2013. The IUCN Red List of Threatened Species. Version 2013.2. http://www.iucnredlist.
org. Downloaded 25 February 2014.
Jones, W. 1824. Discourses delivered before the Asiatic Society: and miscellaneous papers, on
the religion, poetry, literature, etc., of the nations of India. C. S. Arnold. p.28.
Lewis, M. P. & G. F. Simons. 2010. Assessing endangerment: Expanding Fishman’s GIDS. Revue
Roumaine de Linguistique, 55 (2), 103-120
Lewis, M.P., G.F. Simons & C.D. Fennig (eds.). 2013. Ethnologue: Languages of the World,
Seventeenth edition. SIL International, Dallas, Texas. Online version: http://www.ethnologue.com.
Loh, J., R.E. Green, T. Ricketts, J. Lamoreux, M. Jenkins, V. Kapos & J. Randers. 2005.
The Living Planet Index: using species population time series to track trends in biodiversity.
Phil. Trans. R. Soc. B, 360, 289-295.
Loh, J. & D. Harmon. 2005. A global index of biocultural diversity. Ecological Indicators, 5, 231-241.
MacAulay, D. 1992. The Celtic Languages. Cambridge University Press, Cambridge.
Mace, R. & M. Pagel. 1995. A latitudinal gradient in the density of human languages in North
America. Proc. R. Soc. Lond. B, 261, 117-121.
Moore, J.L., L. Manne, T. Brooks, N.D. Burgess, R. Davies, C. Rahbek, P. Williams & A. Balmford.
2002. The distribution of cultural and biological diversity in Africa. Proc. R. Soc. Lond. B, 269,
Nettle, D. 1999. Linguistic Diversity. Oxford University Press, Oxford.
Nettle, D. & S. Romaine. 2000. Vanishing Voices: The Extinction of the World’s Languages.
Oxford University Press, Oxford.
OED Online. Oxford University Press, March 2014.
Oppenheimer, S. 2004. Out of Africa’s Eden. The peopling of the world. Jonathan Ball, Jeppestown.
Pagel, M., Q.D. Atkinson, A.S. Calude & A. Meade. 2013. Ultraconserved words point to deep
language ancestry across Eurasia. Proc. Nat. Ac. Sci., 110 (21), 8471-8476.
Pagel, M. & R. Mace. 2004. The cultural wealth of nations. Nature, 428, 275-278.
Pennisi, E. 2004. The First Language? Science, 303, 1319-1320.
Raghavan, M., P. Skoglund, K.E. Graf, M. Metspalu, A. Albrechtsen, I. Moltke, S. Rasmussen,
T.W. Staord Jr, L. Orlando, E. Metspalu, M. Karmin, K. Tambets, S. Rootsi, R. Mägi,
P.F. Campos, E. Balanovska, O. Balanovsky, E. Khusnutdinova, S. Litvinov, L.P. Osipova,
S.A. Fedorova, M.I. Voevoda, M. DeGiorgio, T. Sicheritz-Ponten, S. Brunak, S. Demeshchenko,
T. Kivisild, R. Villems, R. Nielsen, M. Jakobsson & E. Willerslev. 2014. Upper Palaeolithic
Siberian genome reveals dual ancestry of Native Americans. Nature, 505, 87–91.
Randers, J. 2012. 2052 - A Global Forecast for the Next Forty Years. Chelsea Green, Vermont.
Renfrew, C. 1987. Archaeology and Language: the Puzzle of Indo-European Origins. Cape, London.
Stepp, J.R. et al. 2004. Development of a GIS for Global Biocultural Diversity. Policy Matters, 13,
Sutherland, W.J. 2003. Parallel extinction risk and global distribution of languages and species.
Nature, 423, 276-279.
UNESCO. 2010. Atlas of the World’s Languages in Danger. Third edition. (C. Mosely,
Walker, G. 2003. Snowball Earth. Bloomsbury, London.
WWF. 2012. Living Planet Report 2012 (M. Grooten, ed.-in-chief). WWF International, Gland,
Cashinahua girl, a speaker of a Panoan language with around 1,000 speakers, her face painted with dye from huito fruit. Near the Alto Purus
Reserved Zone, Ucayali, Peru.
© André Bärtschi / WWF-Canon
Biocultural Diversity • threateneD species, enDangereD languages
Why we are here
To stop the degradation of the planet’s natural environment and
to build a future in which humans live in harmony with nature.
© 1986 Panda Symbol WWF World Wide Fund For Nature (Formerly World Wildlife Fund)
® “WWF” is a WWF Registered Trademark. WWF, Avenue du Mont-Blanc, 1196 Gland,
Switzerland Tel. +41 22 364 9111; Fax +41 22 364 0332. For contact details and further
information, please visit our international website at www.panda.org