Project

Back to overview

Developing data-driven spatial analysis methods for inferring spatiotemporal evolution from dialect data

Applicant Jeszenszky Péter
Number 175019
Funding scheme Early Postdoc.Mobility
Research institution Graduate School of Human and Environmental Studies Kyoto University
Institution of higher education Institution abroad - IACH
Main discipline Other disciplines of Environmental Sciences
Start/End 01.06.2018 - 31.08.2019
Show all

All Disciplines (2)

Discipline
Other disciplines of Environmental Sciences
German and English languages and literature

Keywords (13)

spatiotemporal analysis; machine learning; spatial analysis; geographic information science; dialectology; language contact; language evolution; apparent-time constructs; diffusion modeling; route inference; route planning; sociolinguistics; syntax

Lay Summary (German)

Lead
Sprachkontaktmöglichkeiten innerhalb einer Sprache sind von der räumlichen Verteilung der Bevölkerung und der sozio-ökonomische Faktoren mit ihren räumlichen Dimensionen stark beeinflusst und variieren je nach Ort. Es ist ein grosser Reichtum an digital verfügbaren sprachlichen Daten vorhanden, der immer stärker zu nimmt ins besonders im Bereich „crowdsourced“ Sprachdaten und Webkorpora. Trotz der Menge an vorhandenen Daten und Beobachtungen ist nicht viel von der Dynamik und dem Wandel dieser Sprachkontakte quantitativ bekannt. Dies ist zum Teil verursacht durch geringe interdisziplinäre Forschung zwischen GIS und den Sprachwissenschaften. Für leistungsfähige und robuste Analysen benötigen Sprachwissenschaftler datengestützte, räumlich explizite und quantitative Methoden. Die Veränderung der Sprache ist aber implizit die Triebfeder für die gegenwärtige sprachliche Gegebenheiten.
Lay summary

Inhalt und Ziel des Forschungsprojekts

Das übergeordnete Ziel des Projekts ist die Entwicklung von Methoden für ein besseres Verstehen des Sprachkontakts und Sprachwandels im Bezug auf die Dialektforschung. Das entwickelte Rahmenwerk dieses Projekts bestimmt den raumzeitlichen Wechsel basierend auf der gegenwärtigen Verteilung der sprachlichen Merkmale unter Verwendung von GIS Verfahren und Maschine Learning. Zu einem verbesserten Verständnis der Tendenzen im Sprachkontakt und Sprachwandel werden (i) die lokale Kombinationen der bedeutendsten geographischen Faktoren berechnet, (ii) die Sprachwechsel der näheren Zukunft durch raumzeitliche Analysen eingeschätzt und (iii) die Routen der Sprachevolution im letzten Jahrtausend durch Reengineering möglicher Routen basierenden auf lückenhaften und verstreuten Bewegungsdaten eingeschätzt.

Wissenschaftlicher und gesellschaftlicher Kontext des Forschungsprojekts

Adäquate und automatische Analysemethoden zusammen mit sprachwissenschaftlichem Fachwissen können dem Zeitaufwand der Datenerhebung, -analyse und -darstellung verkürzen. Das Projekt wird neue und wichtige Ergebnisse für ein besseres Verständnis des Ausbreitungsprozesses von sprachlichen Merkmalen liefern. Das Rahmenwerk sollte flexibel genug sein um es auf weitere Hypothesen mit anderen Datenquellen aus dem Bereich der Geisteswissenschaften wie beispielsweise Daten zur Demographie oder zur Evolution kultureller Gebräuche anzuwenden. Neben neuen Einblicken in den Sprachwandel liefert das Projekt zudem starke methodologische Beiträge zu GIScience.
Direct link to Lay Summary Last update: 15.12.2017

Responsible applicant and co-applicants

Publications

Publication
Examining geographical generalisation of machine learning models in urban analytics through street frontage classification and house price regression
LawStephen, JeszenszkyPéter, YanoKeiji (2021), Examining geographical generalisation of machine learning models in urban analytics through street frontage classification and house price regression, in 11th International Conference on Geographic Information Science, GIScience 2021, September 27–30, 20, Poznan, PolandDagstuhl Publishing, Saarbrücken/Wadern.
Beyond the Shortest Route: A Survey on Quality-Aware Route Navigation for Pedestrians
Siriaraya Panote, Wang Yuanyuan, Zhang Yihong, Wakamiya Shoko, Jeszenszky Peter, Kawai Yukiko, Jatowt Adam (2020), Beyond the Shortest Route: A Survey on Quality-Aware Route Navigation for Pedestrians, in IEEE Access, 8, 135569-135590.
Japanese Lexical Variation Explained by Spatial Contact Patterns
Jeszenszky Péter, Hikosaka Yoshinobu, Imamura Satoshi, Yano Keiji (2019), Japanese Lexical Variation Explained by Spatial Contact Patterns, in ISPRS International Journal of Geo-Information, 8(400), 1-30.
Spatio-Temporal Prediction of Dialectal Variant Usage
Jeszenszky Péter, Siriaraya Panote, Stoeckle Philipp, Jatowt Adam (2019), Spatio-Temporal Prediction of Dialectal Variant Usage, in Proceedings of the 1st International Workshop on Computational Approaches to Historical Language Cha, Florence, ItalyAssociation for Computational Linguistics, Florence, Italy.
Lexical variation in Japanese dialects revisited: Geostatistic and dialectometric analysis
Jeszenszky Péter, Hikosaka Yoshinobu, Yano Keiji (2019), Lexical variation in Japanese dialects revisited: Geostatistic and dialectometric analysis, in Abstracts of the ICA, Tokyo, Japan 1, 1-9, International Cartographic Association, Tokyo, Japan 1, 1-9.
Witnessing Crime through Tweets: A Crime Investigation Tool based on Social Media
Panote Siriaraya, Yihong Zhang, Wang Yuanyuan, Kawai Yukiko, Mittal Mohit, Jeszenszky Péter, Jatowt Adam (2019), Witnessing Crime through Tweets: A Crime Investigation Tool based on Social Media, in 27th ACM SIGSPATIAL International Conference on Advances in GeographicInformationSystems(SIGSPATIAL), Chicago, IL, USAACM, New York, USA.

Collaboration

Group / person Country
Types of collaboration
Alan Turing Institute (Dr. Stephen Law's project)) Great Britain and Northern Ireland (Europe)
- Publication
Strategic Inform. and Communic. R&D Promotion Programme (SCOPE), Grant#:171507010, PI: Adam Jatowt Japan (Asia)
- in-depth/constructive exchanges on approaches, methods or results

Scientific events

Active participation

Title Type of contribution Title of article or contribution Date Place Persons involved
27th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM SIGSPATIAL 2019) Poster Witnessing Crime through Tweets: A Crime Investigation Tool based on Social Media 05.11.2019 Holiday Inn at the Mart Plaza, Chicago, Illinois, United States of America Jeszenszky Péter;
GIScience2018: Tenth International Conference on Geographic Information Science Talk given at a conference Towards the parameterisation and quantification of dialect contact potential 30.08.2019 RMIT, Melboure, Australia Jeszenszky Péter;
International Cartographic Conference Talk given at a conference Lexical variation in Japanese dialects revisited: Geostatistic and dialectometric analysis 15.07.2019 National Museum of Emerging Science and Innovation (Miraikan) and Tokyo International Exchange Ctr.), Japan Jeszenszky Péter;
Art Research Center seminar presentation series (60th event) Individual talk Language Variation and Change: Geographic and Sociodemographic Causes 12.06.2019 Art Research Center, Ritsumeikan University, Kyoto, Japan Jeszenszky Péter;
The Academic Meeting of the Association of Japanese Geographers Talk given at a conference Historical paths of contact and isolation explain lexical variation in Japanese dialects 21.03.2019 Senshu University, Kawasaki, Japan Jeszenszky Péter;
Postgraduate Seminar presentation series at the Department of Geography, Ritsumeikan University Individual talk Linguistic geography and dialectology: Case studies on Swiss German and Japanese 01.11.2018 Department of Geography, Ritsumeikan University, Kyoto, Japan Jeszenszky Péter;


Communication with the public

Communication Title Media Place Year
Media relations: print media, online media Language Variation and Change: Geographic and Sociodemographic Causes Public seminar live streamed on Youtube International 2019

Abstract

The goal of this project is to develop a methodological framework in a step-wise fashion for inferring spatiotemporal change from the current distribution of the phenomena dialect (in this case dialects spoken in Switzerland and Japan). The framework should be flexible enough to be applicable to other data sources and hypotheses from the broader humanities, such as demography data or the evolution of cultural habits. The first work package (WP1) proposes to quantify the impact of different geographical factors (e.g. physical and sociocultural isolation factors) on linguistic relations, using current linguistic data. Language change is implicitly incorporated as the driving force that lead to the current linguistic situation. WP1 will lead to new linguistic insights that constitute the building blocks for WP2 and WP3, where spatiotemporal language change will be the explicit focus.WP2 is a spatiotemporal analysis of language change over relatively short time spans. It aims to contribute to showing evolution paths of languages by comparing different age groups to infer which regions do presently and will in near future undergo the most profound change. Geographic factors as identified in WP1 are hypothesised to have a major impact on short time change of dialects.WP3 will use current dialect data to infer language change over longer time spans, i.e. hundreds of years, posing a fundamental methodological challenge. GIScience approaches applicable to this problem are concerned with reengineering probable routes from incomplete or sparse movement data, but usually focus on small spatial extents and short time spans. Biology and recently also linguistic typology applies phylogenetic approaches for inferring language trees from current language data, but often ignore spatial evolution paths. The goals of WP3 will be to combine the two methodological frameworks , and model dialect change as a phylogenetic sequence of dialects that follows probable evolution paths in space. In addition to new insights in dialect change, WP3 will make a strong methodological contribution to both GIScience and the humanities.
-