15.5 — Languages

There are roughly 7,000 languages spoken in the world.

About half of them have fewer than 10,000 speakers. The largest twenty are spoken by around half of humanity. A language dies, on the commonly cited estimate, every couple of weeks, and linguists project that a large proportion of the current total will not survive the century.

And the distribution is extremely uneven. Papua New Guinea, with around 10 million people, has over 800 languages. Europe, with far more people, has around 250.

The reasons for all of this are geography, history and power, and this chapter covers them.

Where language families come from

Chapter 6.2 established the method with Indo-European: systematic sound correspondences that cannot arise by chance or borrowing.

The major families, by speakers.

Indo-European — around 3.3 billion speakers. Hindi-Urdu, Bengali, Punjabi, Marathi, Spanish, English, Portuguese, Russian, German, French, Persian, and hundreds more (Chapter 6.2).

Sino-Tibetan — Mandarin and the other Chinese languages, Burmese, Tibetan, and the Tibeto-Burman languages of India's northeast.

Niger-Congo — over 1,500 languages across sub-Saharan Africa, including Swahili, Yoruba and Zulu. The Bantu expansion, which spread a related group of languages across a third of a continent from around 3,000 years ago, is one of the largest such expansions known and is traced by linguistics, archaeology and genetics together.

Afro-Asiatic — Arabic, Hebrew, Amharic, Hausa, and ancient Egyptian.

Austronesian — Malay, Indonesian, Tagalog, Javanese, Malagasy and the Polynesian languages, spread by the seafaring expansion of Chapter 3.6 across half the globe from Madagascar to Easter Island. That Malagasy, spoken off the coast of Africa, is most closely related to a language of Borneo is one of the more startling findings in historical linguistics, and it is well established.

Dravidian — Tamil, Telugu, Kannada, Malayalam and around twenty others, almost entirely in southern India, with the isolated Brahui in Balochistan (Chapter 6.2).

Austroasiatic — Vietnamese, Khmer, and the Munda languages of eastern India.

Turkic, Japonic, Koreanic, Uralic and dozens of smaller families.

And isolates — languages with no demonstrated relatives. Basque, in northern Spain and southwestern France, is the famous European case; Korean is sometimes classed as one; and Nihali in Maharashtra is an Indian example.

Why some languages spread

Four mechanisms and they are all about power.

Agriculture. The farming-language dispersal hypothesis holds that farming populations expanded and carried their languages with them (Chapter 3.7). The Bantu expansion, the Austronesian expansion and the spread of Indo-European are all consistent with it.

Conquest. Latin across the Roman empire; Arabic with the caliphates (Chapter 7.1); Spanish and Portuguese across the Americas.

Administration. A language of government is learned by anyone who wants to deal with the government. Persian in India for six centuries (Chapter 6.9); English afterwards (Chapter 6.19).

And trade. Swahili on the East African coast (Chapter 7.8); Malay across the archipelago; and the various pidgins and creoles.

A pidgin is a simplified contact language with no native speakers, developed between groups with no common tongue. A creole is what it becomes when children grow up speaking it as a first language — at which point it develops full grammatical complexity within a generation. Creoles are the most direct evidence available that grammatical complexity is generated by human minds acquiring language rather than only inherited from a parent language.

Language death

A language dies when the last speaker dies, and long before that it stops being transmitted to children — which is the point of no return.

The mechanism is almost always economic. Parents stop teaching a language to children when it carries no advantage and carries stigma. Speaking the dominant language means school, work, government and dignity.

And it has frequently been enforced. Residential schools in Canada, the United States and Australia removed indigenous children from families and punished them for speaking their languages, with documented and severe consequences. Welsh, Irish, Basque and Breton were suppressed in schools within living memory.

What is lost when a language dies.

Knowledge that is not written down. Names and uses of local plants and animals, navigation, weather signs, and medicinal practice — Chapter 3.3's point that traditional knowledge encodes things nobody has independently verified. Ethnobotanical work has repeatedly found pharmacologically active compounds in plants identified through such knowledge.

A grammatical system. Languages differ in what they oblige a speaker to specify. Some require evidentiality — a marker on every statement indicating whether you saw it, heard it, or infer it. Guugu Yimithirr in Australia uses absolute directions rather than left and right, so speakers must know which way is north at all times, and they do. Each such system is a data point about what human cognition can do, and there are only so many of them.

And an identity. For the community, it is the loss of a link to its own past.

Revival is possible and it is rare and hard. Hebrew is the only substantial case of a language with no native speakers becoming a first language again, achieved by a deliberate movement and then by state policy in Israel. Welsh, Māori, Hawaiian and Irish have been stabilised or partly revived by sustained investment in schooling and broadcasting, which is what it takes.

India

One of the most complex linguistic situations on Earth, and it is worth stating precisely.

The 2011 census recorded 121 languages with more than 10,000 speakers and over 1,300 mother tongues.

Four families are present — Indo-Aryan, Dravidian, Austroasiatic (Munda) and Tibeto-Burman — which is unusual, and it makes India one of the few countries where languages of entirely unrelated families are spoken as neighbours.

And centuries of contact have produced convergence. South Asia is a recognised linguistic area — a Sprachbund — in which unrelated languages have come to share features through long contact: retroflex consonants, the subject-object-verb word order, and certain verb constructions appear across the Indo-Aryan and Dravidian families and are not inherited from a common ancestor. **This is direct evidence of very long coexistence, and it is one of the arguments used in the discussions of Chapter 6.2.

The Constitution's arrangement (Chapter 13.3). The Eighth Schedule lists 22 scheduled languages. Hindi in Devanagari is the official language of the Union; English continues as an associate official language.

And the reason English continues is a specific piece of history. The Constitution provided for English to be phased out after fifteen years, in 1965. When that approached, anti-Hindi agitations in Tamil Nadu were severe — protests, self-immolations and deaths — and the Official Languages Act was amended to allow English to continue indefinitely. The agitation transformed Tamil Nadu's politics and the Dravidian parties have governed the state ever since.

The three-language formula — mother tongue, Hindi and English in Hindi-speaking states; mother tongue, English and another Indian language elsewhere — has been official policy since 1968 and has been implemented very unevenly, and it is currently contested again.

And linguistic reorganisation of states in 1956 (Chapter 6.25) is the settlement that made the arrangement work. Nehru opposed it and the outcome suggests he was wrong: giving language groups their own states removed the grievance rather than feeding it.

Hindi and Urdu

Worth a section because the relationship is genuinely unusual.

They share a grammar and a core vocabulary. Spoken colloquially, they are substantially the same language — a Delhi speaker and a Lahore speaker converse without difficulty.

They differ in script — Devanagari and the Perso-Arabic Nastaliq — and in learned vocabulary, with Hindi drawing on Sanskrit and Urdu on Persian and Arabic.

Their common ancestor is the Hindustani that developed in and around Delhi from the contact between Persian and local speech (Chapter 6.9).

And the separation is largely a nineteenth and twentieth-century political process, in which script and vocabulary became markers of communal identity, accelerating with the events of Chapter 6.23.

The clearest evidence that they remain one language in practice is Hindi cinema, whose dialogue is Hindustani and which is understood across both countries without translation.

English as a world language

Around 1.5 billion speakers, the large majority of whom learned it as a second language.

Why it spread. The British empire first; then American economic, technological and cultural dominance; then the internet, where it was the default early and network effects took over (Chapter 9.11).

It is not because English is easy. Its spelling is notoriously irregular, its phrasal verbs are difficult, and it has a very large vocabulary from having absorbed from everywhere.

Its actual advantages are minor and real: no grammatical gender, very little case marking, and simple verb conjugation compared with most European languages.

And India's position is unusual. English is a genuine Indian language with a large body of literature, a distinct standard variety, and around 10 to 15 percent of the population with some competence — which makes India one of the largest English-using countries in the world while the great majority of Indians do not speak it.

Its effects are both. It is a link language across the north-south divide (Chapter 6.19), a route into international work, and a substantial contributor to the IT industry's success (Chapter 9.14). And it is a class divider that maps almost exactly onto access to education, which is one of the sharpest inequalities in Indian life.

Where this shows up in your life

Every word you have read is Indo-European, most of it Germanic with a heavy Latin and French overlay from 1066, plus Greek for technical terms and a scattering of Arabic, Hindi, Malay and everything else the language has picked up.

And the finding worth carrying. A language is not a fixed object. Every one is a snapshot of a continuous process, changing at a rate that becomes visible over a century. The English of Shakespeare is difficult; the English of Chaucer requires training; the English of Beowulf is a foreign language.

Which means that arguments about correct usage are arguments about which snapshot to privilege. Chapter 8.2 showed printing freezing spellings; the standard varieties of most languages are the dialect of whoever ran the printing presses and the schools.

What the next page covers

Chapter 15.6 closes this Part with food — where the world's staple crops actually came from, how the Columbian Exchange reassembled every cuisine on Earth, why spices mattered enough to launch the voyages of Chapter 8.4, what the industrialisation of food did, and the honest state of the evidence on what people should actually eat.