Paul Heggarty is a senior scientist in the Department of Linguistics at Mac Planck Institute For Evolutionary Anthropology in Leipzig, Germany.

Abstract: From the Tower of Babel to the tales of the Aboriginal Dreamtime, we have long sought to account for our baffling multiplicity of tongues. Linguistic science itself was born out of this curiosity — and by now can look to our language diversity no longer as just an enigma to be solved, but also as a rich seam of data on the human past. Or at least potentially so, for it remains a challenge to work out exactly what the linguistic record really tells us. Certainly, it cannot safely be read without the complementary perspectives of our other windows on the past. This talk surveys how linguistics might both enrich and learn from all its sister disciplines within the original ‘four field’ foundations of anthropology, not least archaeology and population genetics. In their data-sets, methods and analyses these disciplines all differ radically, but they are only all the more complementary for it, towards their common goal of uncovering what is, after all, the same, single human past. Early attempts were bedevilled by false analogies and simplistic associations between languages, ‘cultures’ and ‘peoples’, engendering a generalised distrust of speculations on grand cross-disciplinary synopses. So I return here to first principles, to reconsider how it is that language can inform us of the past at all. Language is a ‘social animal’; it does not ‘just happen’ that language lineages spread, interact, diverge or converge, as if in a social, cultural and demographic vacuum. Rather, those outcomes are but the linguistic reflexes of processes at work far more generally, created by and acting upon the people and societies that speak those languages. These same processes leave their imprints in material culture and the bio-archaeological and genetic records too, so it is here that the link between our disciplines lies. To be convincing and coherent, our respective scenarios need to match independently on three key levels: when, where and why. On each, I survey the main models and methods (both traditional and new) devised for setting our language lineages into their (pre)historical contexts. I assess proposed techniques for linguistic dating, and for locating the ancestral homelands of great language families. And on causation, I peel back the distorting impacts of the modern world to explore the past roles of technologies, trade and cultural networks, state organisation, conquest, subsistence regimes, demography and environment. Any such real-world cause must also be commensurate in scale with whatever linguistic effect it is invoked to explain. Illustrations are taken from across the millennia and across the world, from individual case-studies to the broadest patterns and contrasts in the global linguistic panorama: hotspots and ‘deserts’ of language diversity; powerful convergence areas; and vast, deep-time language families such as Indo-European or Afro-Asiatic, driven by some great expansive and divergent processes, but which? I conclude with the most ambitious generalisations of all, and the furore surrounding them: farming/language dispersals, and the claims to reduce vast proportions of all human languages into great macro-families such as the putative ‘Amerind’ of the New World.