A new approach to language mapping

Mapping languages is a unique challenge, because they have certain characteristics which make it difficult to represent them as a static overlay on a map of the landscape. In this post for Mother Language Day, we’ll look at some of these characteristics and consider how different types of language maps attempt to record them.

Languages are often characterised as complex communication systems which are spoken by a monolingual community spread over a continuous geographical region. An interesting alternative view is to think of languages as communities of practice, like sports teams or book clubs. Under this view, languages are the products of loose associations of individuals who develop certain shared ways of doing and talking about things. Crucially, however, individuals who participate in such communities may retain certain ways of talking which don’t fit with the community’s standards, and they may also participate in other communities of practice and adopt and be influenced by their standards. If we accept this conception of languages, the challenge of mapping them becomes clear: how do you define the boundaries of a language, both socially and geographically, when individuals may move freely between different places and speak different languages, dialects and registers, even in the course of a single day?

The answer to this question will not be a simple one, and yet it’s a challenge worth addressing. A good language map should be a comprehensive record of the geographical spread of the language’s community of practice. It should also take into account the fact that certain places may be occupied simultaneously by other communities of practice, and that individuals within each community may also participate in other communities, either in that place or elsewhere. To put it more simply, a good language map should record the complex relationships between people, places and languages.

Traditional methods of language mapping cannot provide this kind of record, partly because of their static nature and partly because of the uneven numbers of speakers of each language. For example, in the map below, some major European languages are shown as coloured polygons representing their supposed geographical spread over the continent:


From Wikimedia Commons

While this map gives us a good general idea of the approximate locations of the major languages, it is woefully inaccurate as a record of the actual spread of languages in Europe. It does not record many smaller languages or varieties of larger languages, it does not record the movement of people around the continent, and it does not record individuals and their relationship to the various language communities that they participate in. Geographical language boundaries in this type of map are also depicted as similar to political boundaries, with only some very general overlaps recorded with the use of cross hatching.

It is perhaps a bit unfair to pick on such a broad scale map, as some distortions are necessary in order to keep it relatively uncluttered. But the problems inherent in traditional methods of language mapping are made clear by this example. Similar problems also occur when points are used to represent entire languages. Where, for example, should we place the point for English given its global distribution?

In order to make better language maps, we need to develop a more sophisticated approach. This can be achieved by modelling the relationships between people, places and languages in a more fine-grained way, and thinking of a different approach to recording these relationships. The method we have come up with for Language Landscape is to use audio and video recordings as the basic point of intersection between person, place and language. As long as the person who made the recording collected accurate information at the time, with a recording we can be sure that at that time, in that place, that individual said something in that language. The recording is therefore a small piece of evidence about the relationship of a language to a place and to the individual that spoke it. Mapping that recording along with information about the language and speaker is therefore one contribution to the map of that language. Repeating the process with every person who speaks that language would result in a comprehensive map of the language.

Of course, in most cases, persuading all speakers of a language to contribute to such a map would be unfeasible. This is especially true for larger languages, though with smaller communities of tens or hundreds of individuals it may well be possible. The difficulty of persuading people to map their recordings is the principal reason why traditional language maps like the one above still have their place, as they provide useful information without the need for lots of people to contribute. The problems inherent in them still stand, however, and it is intriguing to imagine what we might discover about the complex relationships between languages, people and places if more people contributed recordings to the Language Landscape map. So on Mother Language Day, we encourage you to represent your mother languages on Language Landscape and help to build a better record of all the places they are spoken around the world.