Project

Back to overview

High Mountains Low Arousal? Distant Reading Topographies of Sentiment in German Swiss Novels in the early 20th Century

English title High Mountains Low Arousal? Distant Reading Topographies of Sentiment in German Swiss Novels in the early 20th Century
Applicant Herrmann Berenike
Number 189832
Funding scheme COST (European Cooperation in Science and Technology)
Research institution Digital Humanities Lab Philosophisch-Historische Fakultät Universität Basel
Institution of higher education University of Basel - BS
Main discipline German and English languages and literature
Start/End 01.06.2020 - 31.05.2023
Approved amount 239'981.00
Show all

All Disciplines (5)

Discipline
German and English languages and literature
Psychology
Political science
Swiss history
Social geography and ecology

Keywords (14)

literary history; emotion; gender; fictional topographies; cultural heritage computing; distant reading; text mining; digital corpora; European novel; literary characters; sentiment analysis; Swiss literature; geography of literature; digital humanities

Lay Summary (German)

Lead
Das Forschungsprojekt «High Mountains Low Arousal? Distant Reading Topographies of Sentiment in German Swiss Novels in the early 20th Century» verfolgt einen innovativen Ansatz der Digitalen Literaturgeschichtsschreibung. Unter der Leitung von Dr. J. Berenike Herrmann werden am Digital Humanities Lab der Universität Basel datenwissenschaftliche Methoden (Sentiment Analysis, Named Entity Recognition) angewandt, um die Verbindungen von Raum- und Gefühlsdarstellungen in einem umfänglichen Korpus Schweizer Romane aus der Zeit 1900 -1930 zu untersuchen. Diese können auch in Bezug auf weitere deutschsprachige Literaturen und den Europäischen Kontext ganz neu untersucht werden.
Lay summary

Mittels des sogenannten «Distant Reading» sind in der Literaturwissenschaft neuerdings statistische und visuelle Analysen von Emotion und Stimmung, aber auch von fiktionalen Raum- und Geodaten möglich. Unser Projekt nutzt diese Möglichkeiten, um die Umrisse einer «fiktionalen Deutschschweizer Emotionstopographie» zu zeichnen.

Das Kernstück des Projekts ist ein digitalisiertes Literaturkorpus, das u.a. in Zusammenarbeit mit der UB Basel und der internationalen COST-Action "Distant Reading the European Novel" erstellt wird. Durch die digitale Modellierung hunderter Romane, die die deutschschweizer, österreichische und deutsche Erzählliteratur repräsentativ abbilden, unterscheidet sich das Projekt von der klassischen Literaturgeschichte schon durch seine Skalierung.

Das Projekt entwickelt zudem eine spezifisch literaturwissenschaftliche Methodik der automatischen Emotions- und Stimmungsanalyse, aufbauend auf Tools aus der digitalen Marktforschung und der Geo-Informatik. Die bestehenden Tools werden u.a. in Zusammenarbeit mit Emotionspsychologen (Prof. Arthur Jacobs, FU Berlin) so adaptiert, dass literarische Darstellungen von Gefühlen in Büchern der Jahrhundertwende von 1900 messbar und digital auf räumliche Koordinaten der fiktionalen Welten projiziert werden können.  

Forschungsfragen zur Spezifität der deutsch-schweizerischen Literatur sollen damit neu aufgerollt werden, etwa bezüglich eines vergleichsweise stärker ausgeprägten «Realismus» der deutschschweizer Literatur in der Moderne, oder auch nach Beziehung bzw. Differenz zu einer «Heimatliteratur»-Schilderung von (alpinem) Raum. In einer weiteren Studie sollen die Sentiment-Graphen der fiktionalen Ereignisse verglichen werden: Was sind dominante Muster? Wie stark sind jeweils die «Gefühlsamplituden»? Fragen nach einer «deutschschweizer» Literatur sollen so im Spannungsfeld von Nationalliteratur, lokalen Bezügen und einem pan-europäischen Kosmopolitanismus digital neu erforscht werden.

Direct link to Lay Summary Last update: 11.05.2020

Responsible applicant and co-applicants

Employees

Project partner

Associated projects

Number Title Start Funding scheme
183221 Forschungslogiken in den textbasierten Digital Humanities: Analysen von Bewertungspraktiken nach dem Machine Learning-Turn 01.12.2018 Digital Lives

Abstract

The project High Mountains Low Arousal? - in close collaboration with the international COST Action “Distant Reading the European Novel” - proposes a contribution to literary studies in terms of digital resource building, literary theory, and methodological development. By means of a distant reading of German-Swiss novels, it aims at pioneering comparative historiographical and systematic research on the German novel of the early 20th Century, using digital methodologies of advanced sentiment analysis (SA). Following the recent examples of theoretically sophisticated distant reading (Piper, 2018a; Underwood, 2019), the proposed project will focus on narrative patterns tracing emotional arcs (Reagan, Mitchell, Kiley, Danforth, & Dodds, 2016) and on the emotional encoding of fictional space and literary characters.Sentiment analysis is one of the key methodologies of distant reading, yet its valid domain adaptation to German literary discourse is still pending. As SA is language-specific and highly sensitive to historical and discursive variation, its advanced domain adaptation for literary analysis is costly, but offers a strong gain in lending an unprecedented systematicity to exploration, hypothesis testing and prediction of literary forms and patterns. My approach applies a psychologically validated conceptualization of emotion, combining a dimensional approach to sentiment valence (positive, neutral, or negative sentiment) and arousal (strength of sentiment and emotion), with one to discrete emotions (such as fear, anger, joy).One of the main contributions of the proposed project is the development of the German-Swiss Novel Corpus (SNoC), covering the period 1900-1930. It will be a unique resource specifically built for the project as an XML database for full texts and meta-data, retro-digitizing 100 German Swiss novels and harvesting around 150 further ones; it will be complemented by the German Novel Reference Corpus (GNoC), that will be compiled from existing repositories, with ca. 600 novels written by non-Swiss German authors. Meta-data will be harvested using automatic and manual techniques, identifying sub-genres, canonicity, and information on author gender and ideology. The project aims at contributing a new data-driven and philologically reflected perspective on the historiography of Swiss novels of the early 20th Century, situated within the transnational continuum of the German-speaking countries. It will critically examine questions of the specificity of Swiss novels as compared to German and Austrian ones, addressing a putative Swiss marginalization, mediocracy, and predominance of anti-modernist realism surfacing in idyllic landscapes and simple plots. Through thus, it will shed a new view on ideas of Swissness transported by novels during the period, in between educative mediation (Helvetia mediatrix, Lecoultre) and the enactment of a national myth unfolding in spatial topographies that center around the Alps.Offering the first distant reading of the Swiss novel applying SA, the project is comparative within Swiss literature as well as at a transnational level. The overarching aim is to offer a data-driven exploration of the patterns of specificity of Swiss literature in close collaboration with the CA “Distant Reading the European Novel” - to advance digital resource building, literary theory, and methodological development. In all this, the project will build on the multi-disciplinary expertise of the COST Action “Distant Reading for European Literary History”, but will go beyond its general scope in applying a more fine-grained text analysis. The project's goals - (1) building a fine-grained representative corpus of Swiss, as well as German and Austrian, novels of the period 1910-1930, (2) adapting and developing tools of sentiment and emotion analysis for German literary discourse of the period, and (3) furthering insight into the representation of emotion and sentiment in Swiss novelistic writing during a period of European crisis and cultural advancement - will match, but also surpass, specific work packages of WG1 (https://www.distant-reading.net/wg-1/), WG2 (https://www.distant-reading.net/wg-2/), and WG3 (https://www.distant-reading.net/wg-3/). The project will contribute to the re-evaluation and redefinition of key concepts for literary history before the background of applying quantitative methodology and explore their implications for periodization and canonization in European literary history.
-