Johnson: Mountains high enough and rivers wide enough
The physical landscape seems to affect the diversity of language
By R.L.G. | BERLIN
WHY do some places in the world have lots of small languages, and others have fewer, bigger languages? Earlier studies seemed to show that areas of high altitude, rainfall and temperature had high cultural and linguistic diversity. A brief glance in the direction of the geography and linguistic diversity of the Caucasus, central Africa or New Guinea (pictured) would seem to bear this out.
A new study has narrowed in on the details, and found that two features of the landscape predict language diversity all around the world. Jacob Bock Axelsen and Susanna Manrubia, in the Proceedings of the Royal Society B, looked at a number of environmental factors including vegetation, temperature, precipitation, altitude, landscape “roughness”, density of rivers, distance from lakes and population density. They carved up the world into squares 222km wide and looked only at the 100 biggest land masses (to correct for the fact that languages spoken on small islands are less likely to spread). They also used some clever statistical work to account for the fact that some of these factors (like rainfall and vegetation) are themselves partially correlated.
The result: river density and landscape roughness are correlated with language density in all the world’s regions. (“Roughness” is the extremity of topographical variation; a high plateau is not “rough”, but a low yet jagged mountain chain is.) None of the other factors was significant in all four of the regions: Africa, “extended” Asia (including Oceania and Australia), the Americas and Europe. The significant variables were most significant in Africa.
The study is interesting in many wider ways. The interplay of physical geography and the rise of cultures is a fascinating one, explored famously in Jared Diamond’s book “Guns, Germs and Steel”. The reason some societies become populous and powerful may have a lot to do with the raw materials around them. Mr Diamond, for example, focused on plants that can be cultivated and animals that can be domesticated, leading to stable food supplies, division of labour and so on.
Ms Manrubia’s and Mr Axelsen’s study looks at the land itself. They hypothesise sensibly that rivers and mountains both tend to form natural barriers. These might tend to keep smaller cultures from mingling and fusing into bigger ones. Such physical barriers are also associated with speciation, as populations separate and evolve into new species. (Darwin himself was inspired by linguistic evolution in hypothesising the biological kind.) Rivers, of course, also provide transport; the authors speculate that where rivers meet, cultures tend to do so as well, leading to new “contact” languages resulting from (originally imperfect) communication between the two groups. (On the biological analogy, this would be something akin to genetic recombination.)
Significantly, Africa and Asia turn out most similar to one another in the study. In addition to rivers and roughness, average temperature and precipitation were also positively correlated with language density there. The authors hazard, plausibly enough, that the two regions represent something like an earlier state with regard to the land-to-language link. The rise of powerful states in Europe, and those states’ subsequent rapid colonisation of the Americas, led to the loss of huge numbers of languages in both regions. In the Americas, just four languages (English, French, Spanish and Portuguese) dominate the land from Alaska to Tierra del Fuego, with native languages largely confined to high-altitude redoubts. Europe, for its part, is the most atypical region with regard to the predictive power of rivers and roughness. So these factors are significant, but others will of course have co-determined the rise of larger languages and cultures in Europe (and elsewhere).
Though not the main focus of their paper, Ms Manrubia and Mr Axelsen play with their data by applying the landscape-to-language fits of Europe and America to Africa and Asia, asking in effect, “What if the language-diversity development of Africa and Asia were to follow the path of Europe and the Americas?” The resulting “predicted” language density would see 3,700 of the world’s languages disappearing from Africa and Asia—startlingly close to the prediction that half of the world’s 7,000 languages might die out in the next century.
On this map produced by Ms Manrubia and Mr Axelsen, areas that are dominated by widespread languages or uninhabited areas are indicated in green. Other colours have been assigned randomly; each one represents a different language.
