The DIACL/TITUS Polygon Archive is an open resource, which has a long history of compilation. Ultimately, the archive comes from two sources, the DiACL database and TITUS. The interface for accessing the Polygon Archive is inside the DiACL database and therefore, polygons are linked to languages and language metadata in DiACL. This interface for uploading, displaying and downloading polygons in json and txt format, either as full or simplified, was already previously built in DiACL. However, at the time of the transfer of DiACL to Goethe University, there were only around 150 polygons in DiACL. A majority of the polygons in the Polygon Archive stem from the TITUS database, which compiled language polygons for almost all of the world (except for Australia and South East Asia) already in the 1990s. These polygons have been georeferenced, digitized and uploaded to the DiACL polygon interface, building the new Polygon Archive.
How to use the Polygon Archive
On the webpage, existing polygons are given on a map (new polygons are continuously added). In this map, there are overlapping polygons, which means that polygons may be hidden behind other polygons. In general, languages with smaller area (dialects, minority languages) are on top of languages with larger areas, but there are also polygons from ancient languages and dialects. This means that the most convenient way to look for polygons is either to download all polygons or to search by Language.
There are over 3000 languages in DiACL, but at current state only 800 polygons, which means that not all languages have polygons.
Language polygons can be very complicated, in particular for languages that have changed their extension dramatically in recent centuries. Different from many other similar resources, the Polygon Archive gives a precolonial extension of languages, meaning that English is spoken mainly on the British Isles, French in France, and Spanish in Spain.
Another complicating issue is that indigenous languages of hunter-gatherers, pastoralists, or horticulturalist populations frequently move around due to their subsistence system. The maps in TITUS, which stem from various geographic sources, often target a language continuum, or a language node rather than a language or dialect. For the purpose of using these polygons, we have introduced the language reliability type Continuum, which typically corresponds to a node in Glottolog trees (these nodes have Glottocodes, and are labelled as “Languoids”). An example would be Na-Dené, in Glottolog known as Athapaskan-Eyak-Tlingit, with a code atha1245, which is by definition a language family. In the future, polygons like these could potentially be refined and improved to contain a higher number of daughter languages or dialects in polygon form.