LanGIS Talks: A lecture series on spatial analysis and GIS for language study, autumn 2018

The talks take place at Engelska parken, between 10:15 and 12:00.

November 20 (Tuesday)
Room: 16-0054

Terhi Honkola (The BEDLAN project at Turku University)


Landscape genetics of dialects: Barriers and pathways structuring the spatial pattern of linguistic variation

Geographical distance is a common factor separating populations and inducing population genetic structure. Pure geographical distance may, however, only be a coarse estimation of the true landscape connectivity between areas as various landscape features, both natural and anthropogenic, may either promote or limit the actual connectivity. Similarly to populations and their genes, geographical distance and landscape connectivity may shape the spatial pattern of linguistic variation. Here we study which natural and/or anthropogenic barriers and pathways have played a role in shaping the dialect division of Finnish. We studied this question with a ca. 100 years old dialect data of Finnish, and with landscape variables such as water routes, topography, eskers, watersheds and roads in the 16th century. The landscape variables were digitized for the whole area of Finland and turned as speed of travel surfaces. We studied the effect of individual variables and their combination on linguistic differences with distance-based statistical methods. We studied both early (south and central Finland) and later (whole Finland) stages of the spread of the Finnish language. According to our results, a combination variable of different environmental features explained the linguistic differences in the southern and central Finland, while eskers explained that of the whole Finland. This suggests that landscape features have shaped the spatial pattern of linguistic variation of Finnish.

October 16 (Tuesday)
Room: 6-0022

Peder Dam (Odense City Museums)


Large and classified GIS data in the humanities: Creation, classification, mapping and analysis

Within the humanities, GIS is already widely used to store and visualize large amounts of data, as it is common to use GIS for analyzing local areas and smaller regions. Nationwide or similar large mappings and analyzes of classified GIS datasets are, on the other hand, more rare in the humanities. This is probably due to many reasons. Other sciences have a longer tradition for approaching research questions through large and statistical data. Furthermore, is humanistic data often difficult to put into a database because of the complexity of the information. And last, but not least, is it time consuming to produce large and well-qualified GIS data.

However, the advantages and perspectives of large and classified GIS datasets are obvious. It can provide insight, overview and a more systematic understanding of geographical variations, which is more difficult to achieve through more traditional case-oriented studies. At the same time, due to their geographical location, the data can be compared analytically with other humanistic datasets or with datasets from the other sciences.

The paper is based on my work with humanistic GIS data over the past 20 years, but with particular focus on the creation and analysis of a dataset containing information from 15,000 historical Danish settlements and their place names from my PostDoc project Bebyggelser og stednavnetyper. The book (2015) is sold out, but the publisher has published it as Open Access:

October 9 (Tuesday)
Room: 16-0054
Ljuba Veselinova (Department of Linguistics, Stockholm University)


GISMaps and Mapping in Linguistic Research

In this talk I introduce basic concepts of Geographical Information Systems (GIS) and their applicability to linguistic research. I also offer a critical overview of the ways language mapping has been traditionally done following (Dahl and Veselinova 2005).

I argue that despite the fact that the linguistic community has been mastering the GIS technology in the past 15 years, we continue to be very traditional in our approach to showing the location of the languages in the world. For instance, the language composition of urban areas is still largely understudied. We also remain unaware of geographic definitions and standards with regard to GIS workflow and cartography. As a partial illustration of this, I discuss two existing projects, Langscape, University of Maryland,  and the UNESCO Atlas of the World’s Languages in Danger Both of them can be seen as innovative in the sense that they offer dynamic language maps. At the same time, all of these projects show, in varying degrees, a more or less traditional approach to language mapping.

At present, geo-linguistic data typically come in two forms, (i) maps where language location is represented by means of abstract areas and (ii) listings of populated places and areas in language descriptions. As pointed out by (Briscoe, 2009), such sources do not stand the scrutiny of data collection process from a GIS perspective. Specifically, typical attributes of GIS data such as spatial accuracy, data completeness as well as documentation of the collection process are at best scarce, if present at all. However, since the use of GIS tools in linguistic research is still in its infancy and a lot of valuable information is found in the traditional sources just mentioned, they cannot be lightly discarded.

Finally, I argue that when adopting the GIS technology, we should depart from our traditional view of maps as picture products but rather conceive of them, first and foremost, as visualizations of data. That is, when designing maps, we should adopt the geographer’s perspective on mapping and cartography as much as possible. It is also essential to define our own standards for language mapping and include teaching of relevant subjects in our programs.


Briscoe, Ulla. 2009. Geolinguistic GIS Applications: Aspect of Data Quality in Mapping Lesser-Used Languages, Centre for Geoinformatics (Z_GIS), Salzburg University.

Dahl, Östen, and Veselinova, Ljuba. 2005. Language Map Server. Paper presented at 25th ESRI International User Conference, San Diego.

September 25 (Tuesday)
Room: 16-2041

Olof Karsvall (The National Archives of Sweden), PhD in Agrarian History, M.S. in Computer Science, B.Soc.Sc in Human Geography, TORA project leader


TORA - Linking historical sources to places on the semantic web

TORA is a historical geographic register provided by the Swedish National Archives. It uses a cloud service, where information about people, events and things in various historical documents and datasets could be linked to the places defined in TORA. Currently about 50 percent of all villages and farms in medieval and early modern Sweden are covered. Also, historical divisions, for instance parishes, are included. Using linked data technologies, the goal is to increase, simplify and improve the use of historical sources in research and applications. In this presentation, I will share experiences from working with linked data and highlight some directions within "digital" historical research.

TORA has been initiated by a group of researchers and co-workers at several institutions in Sweden, with funding (2015–20199) from Kungl. Vitterhetsakademien, Riksbankens Jubileumsfond and the Swedish National Archives. A first public launch will take place 2018–2019.

September 18 (Tuesday)
Room: 2-1024

Jonah Åhlfeldt (Lund University)


Geospatial and temporal visualization of cultural heritage data: towards a collaborative platform for datadriven historical information in Swedish research

In Humanities research, there is a need for a collaborative and sustainable platform based on the principles of Linked Open Data to participate in real time data exchange with already existing external datasets. Through a set of already existing online databases, the presenter will describe the core building blocks and the challenges such a platform will face. All these projects are data-driven historical applications which have geospatial and temporal information at their core. They also depend heavily on the use of already available external data. Furthermore, they all use interactive maps as the primary visualization of data. The Digital Atlas of the Roman Empire, hosted at Lund University, has been part of the Pelagios Commons collaboration since 2012, providing both a digital map and a gazetteer of ancient places for everyone to use, and at the same time, using data produced by other parties in the network. The basis of data integration and exchange is the use of common authority data for ancient place. In the same way, the Diabas - North Germanic geolexical database, Umeå University will rely on authority data for parish to be able to use external resources from the cultural heritage sector, and at the same time provide data to others. Furthermore, the Key to Uppåkra online database, Lund University, will use data from the Swedish heritage board to provide archaeological context to the archaeological site of Uppåkra. A collaborative platform will enable data exchange by means of stable identifiers of data items (URI:s), the use of common authority data, serialization of data in different formats, selection and automatic export of data, and mechanisms for the retrieval of external data in real time. The platform will be able to contribute to both international and national collaborations and at the same time benefit from already openly available data.