More than 4,000 years before the recent expansion of Europeans over all other continents, there was an earlier expansion within Europe and western Asia that sired most of the languages spoken in that region today. Although those earlier conquerors were illiterate, much of their language and culture can be reconstructed from shared word roots preserved in modern Indo-European languages. Their conquest of much of Eurasia, like the subsequent overseas expansion of their descendants, appears to have been an accident of biogeography. 'Yksi, kaksi, kolme, nelja, viisi.

I watched the little girl counting out five marbles, one by one. Her was familiar, but her words were strange. Almost anywhere else Europe, I would have heard words like our English 'one, two, three' 'uno, due, tre' in Italy, 'ein, zwei, drei' in Germany, 'odin, dva, tri' in Russia. But I was vacationing in Finland, and Finnish is one of Europe's) non-Indo-European languages.

Today, most European languages and many Asian languages as far as India are very similar to each other (see table of vocabulary overle; No matter how we complain while memorizing French word lists school, these so-called 'Indo-European' languages resemble English, each other, and differ from all the world's other languages, in vocabul and grammar. Only 140 of the modern world's 5,000 tongues belong this language family, but their importance is far out of proportion to tt numbers. Thanks to the global expansion of Europeans since 149! especially of people from England, Spain, Portugal, France, and Russi nearly half the world's present population of five billion now speaks Indo-European language as its native tongue.

To us it may seem perfectly natural, and in no need of further explanation, that most European languages resemble each other, until we go to parts of the world with great linguistic diversity do realize how weird is Europe's homogeneity, and how it cries out explanation. For example, in areas of the New Guinea highlands where I work and where first contact with the outside world began only in the Twentieth Century, languages as different as Chinese is from English replace each other over short distances (Chapter Thirteen). Eurasia must also have been diverse in its pre-first-contact condition, and gradually become less so until finally some people speaking the mother tongue of the Indo-European language family steamrollered almost all other European languages out of existence.

INDO-EUROPEAN LANGUAGES

English one two three mother brother sister

German ein zwei drei Mutter Bruder Schwester

French un deux trois mere frere soeur

Latin unus duo tres mater frater soror

Russian odin dva tri mat' brat sestra

Old Irish oen do tri mathir brathir siur

Tocharia sas wu trey macer procer ser

Lithuania vienas du trys motina brolis seser

Sanskrit eka duva trayas matar bhratar svasar

PIE oynos dwo treyes mater bhrater suesor

NON-INDO-EUROPEAN LANGUAGES

Finnish yksi kaksi kolme aiti veli sisar

Fore ka tara kakaga nano naganto nanona

PIE stands for proto-Indo-European, the reconstructed mother tongue of the first Indo-Europeans. Fore is a language of the New Guinea Highlands. Note that most words are very similar among the Indo-European languages and totally different among the non-Indo-European languages.

Of all the processes by which the modern world lost its earlier linguistic diversity, the Indo-European expansion has been the most important. Its first stage, which long ago carried Indo-European languages over Europe and much of Asia, was followed by a second stage that began in 1492 and carried them to all other continents (Chapter Fourteen). When and where did the steamroller start, and what gave it its power? Why was Europe not overrun instead by speakers of a language related to, say, Finnish or Assyrian?

While the Indo-European problem is the most famous problem of historical linguistics, it is a problem of archaeology and history as well. In the case of those Europeans who carried out the second stage of the Indo-European expansion beginning in 1492, we know not only their vocabulary and grammar but also the ports where they set out, the dates of their sailings, the names of their leaders, and why they succeeded in conquering (Chapter Fourteen). But the quest to understand the first stage is a search for an elusive people whose language and society lie veiled in the pre-literate past, even though they became world conquerors and founded today's dominant societies. That quest is also a great detective story, whose solution depends on a language discovered behind a secret wall in a Buddhist monastery, and on an Italian language inexplicably preserved on the linen wrappings of an Egyptian mummy.

When you first think about it, you might be excused for dismissing the Indo-European problem as obviously insoluble. Since the Indo-European mother tongue arose before the origin of writing, is it not almost by definition impossible to study? Even if we found the skeletons or pottery of the first Indo-Europeans, how would we recognize them? The skeletons and pottery of modern Hungarians, living in the centre of Europe, are as typically European as goulash is typically Hungarian. A future archaeologist excavating a Hungarian city would never guess that Hungarians speak a non-Indo-European language, if no examples of writing itself were recovered. Even if we could somehow locate the place and time of the first Indo-Europeans, how could we hope to deduce what advantage let their language triumph?

Remarkably, it turns out that linguists have been able to extract answers to these questions from the languages themselves. I shall explain why we are so confident that language distributions today reflect a steamroller in the past, then try to assess when and where the mother tongue was spoken, and how it managed to take over so much of the world. How can we infer that the modern Indo-European languages replaced other, now-vanished languages? I am not talking about the observed second-stage replacements of the past 500 years, which saw English and Spanish dislodge most native tongues of the Americas and Australia. Those modern expansions were obviously due to the advantages Europeans gained from guns, germs, iron, and political organization (Chapter Fourteen). Instead, I am talking about the inferred first-stage replacement that saw Indo-European dislodge older languages of Europe and western Asia, and that must have happened before writing reached those areas. The map on the following page shows the distribution of Indo-European language branches surviving in 1492, just before Spanish began to leap across the Atlantic with Columbus. Three branches are especially familiar to most Europeans and Americans: Germanic (including English and German), Italic (including French and Spanish), and Slavic (including Russian), each branch with twelve to sixteen surviving languages and 300 to 500 million speakers. The largest branch, however, is Indo-Iranian, with ninety languages and nearly 700 million speakers from Iran to India (including Romany, the language of gypsies). Relatively tiny surviving branches are Greek, Albanian, Armenian, Baltic (consisting of Lithuanian and Latvian), and Celtic (including Welsh and Gaelic), each with only two to ten million speakers. In addition, at least two Indo-European branches, Anatolian and Tocharian, vanished long ago but are known from extensive preserved writings, while others disappeared with less trace.