Skip to main content

Ancient Greek historians have much to offer readers in the 21st century, but there is an as-yet untapped contribution of great scientific relevance they can make. By way of illustrating this contribution, we report here on results of the “Herodotos Project”, an investigation that draws on Herodotos, and ultimately other Greek historians, to document languages and dialects of antiquity so as to better understand the rate of language death in the past.

It is often said that the rate of language death (by various causes including, but not restricted to, both enforced and voluntary shift from one language to another for economic or social advancement) is greater now than ever before. Without minimizing the urgency such statements are meant to convey, one must realize that they are made without a clear empirical basis, in the form of a calculation of the rate of language loss in pre-modern times against which to measure the current rate. One estimate, by evolutionary biologist Mark Pagel (1995: 6), counts as many as 400,000 languages spoken since the beginning of human language (c. 75,000 B.P.); given estimates of some 7,000 contemporary languages, there has been vast language loss over the millennia.

Still, hard empirical data on this question is a desideratum. To that end, we have been establishing a point of comparison by using information from historical sources that give a basis for estimating how many languages have died within pre-modern recorded history. Specifically, using Herodotos’s Histories, we are determining the number of languages that have not survived in any form; this standard from the past may allow calculation of the rate of language loss in ancient times, a basis for meaningful comparison with the modern rate of language death.

It is of course tricky to extrapolate such linguistic information from early sources but, despite his limitations, Herodotos does provide the most comprehensive potential listing of languages of the ancient world in any Classical source. This project does not simply infer the existence of different languages from Herodotos’s mention of different peoples, both because of the well-known “language” versus “dialect” conundrum, and because he generally offers no substantive linguistic information. Rather, by making reasonable assumptions—based on geography, known linguistic lineages, and later sources and compilations such as Stephanus of Byzantium’s Ethnica—we develop statistics on how many distinctancient languages there are that are mentioned in early sources and how many have survived or have died off.

We report here on language information contained in the first four books of Herodotos and our language-loss calculations from that information. To date, our examination has yielded upwards of 130 distinct ethnonyms, a significant portion of which we expect to constitute linguistically distinct entities. From these figures, we calculate a rate of language loss in ancient times for comparison with modern estimates of 300 languages to be lost per decade throughout the 21st century. We plan to extend this analysis to cover all of Herodotos as well as other ancient historical sources.

Although our calculation is based on many assumptions, mining ancient sources in this way may be the only means available for such a determination. Labov (1994: 7) has famously characterized historical linguistics as the “art of making the most of bad data”: the works of Herodotos and other ancient historians may be bad (i.e., imperfect or flawed) data, but they are the best data we have for understanding language diversity in ancient times and thus for gaining some insight into language loss in pre-modern times. We thus see this use of Herodotos (et al.) as a small step toward answering a big question, namely: is contemporary language loss just part of the natural evolution of languages or are globalization and other modern forces accelerating the rate of loss?