The Evolution of Complex Text Patterns: Development and Application of a corpus-linguistic Approach for Analysis of Diachronic Change of Text Patterns in its Multidimensionality

• Applicants: Privatdozent Dr. Alexander Geyken; Professorin Dr. Britt-Marie Schuster
• Subject Area: Applied Linguistics, Computational Linguistics ; General and Comparative Linguistics, Experimental Linguistics, Typology, Non-European Languages
• Term: from 2019 to 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 417702242

• Final Report Year: 2023

Final Report Abstract

The genesis and evolution of text types serves as an important indicator of the communicative tasks which are relevant for a society, since text types bear within them the commonly accepted patterns for solving such recurring tasks. Based on a multidimensional text model, our project aimed to create a standard workflow for textual analysis (and the analysis of historical texts in particular), in order to gain insights on the development of text types in their shapes and peculiarities (also referred to as text patterns). The shape of a text pattern evolves (1) from common purposes of certain texts and the transformation of such purposes into communicative acts (functional dimension), (2) from common topics and their characteristic development and handling in a text (thematic dimension), (3) from common ways to establish relationships in certain texts, esp. between producer and recipient (relational dimension) and (4) from characteristic ways in which communicative acts are carried out (stylistic dimension). Our studies focused on corpora of two domains, including first, 352 works of 17th and 18th c. devotional literature including devotional prose, books of prayer, and sermons), and second, 174 issues of the newspaper Allgemeine Zeitung published between 1830 and 1929. Texts were either newly digitized or reused from the Deutsches Textarchiv (DTA) corpora and are now collectively available in the Internet. The project was designed to combine close reading of selected texts, manual annotation of larger amounts of text based on linguistic pragmatics and the application of automatic quantifying methods for text analysis. Manual annotation was pursued on newspaper articles as well as sermons and included the creation of elaborate tag sets which were repeatedly validated in the course of annotation. These tag sets together with the manually annotated texts themselves will be publicly made available for subsequent use by the scholarly community. For quantitative text analysis, in addition to applying established techniques of corpus linguistics we developed new methods based on TEI markup to identify significant types of text structures or to evaluate the influence of position to the significance of textual phenomena. Furthermore, methods for the identification of stylistic procedures were developed. The project results are: a) A substantial amount of textual patterns can actually be identified by quantifying procedures. However, this cannot be achieved 'by pushing a button', but only by repeated calibration of the patterns under consideration. b) There are, however, further patterns which can only be identified through close reading and manual annotation. In our detailed studies, which are presented in the collective volume "Historische Textmuster im Wandel: Neue Wege zu ihrer Erschließung" ("Historical Text Patterns in Transition: New Ways of their Detection"; Reihe Germanistische Linguistik 2023), it becomes apparent that in order to gain valid results on textual patterns for solving common communicative tasks, it is necessary to tightly interlock qualitative and quantifying methods. Based on these results, we propose a standard workflow that can be applied to historical as well as modern texts. The longitudinal analyses carried out in the project show that both devotional literature and press texts, especially those of a rhetorical character, change considerably in the periods under consideration. The change in press text types is due, for instance, to the gradual emergence of a new professional profile in journalism, which e.g. becomes apparent through new ways of shaping relationships or the preference for new stylistic patterns. In devotional text types such changing preferences for stylistic patterns may e.g. reflect the variance and alteration of theological traditions.

Publications

• Art und Funktion von typographischen Mitteln zur Textgliederung in erbaulichen Textsorten des 17. Jahrhunderts. Automatische Analyse im Korpusvergleich und qualitative Einordnung. In: Franz Simmler und Galina Baeva (Hrsg.): Textgliederungsprinzipien. Ihre Kennzeichnungsformen und Funktionen vom 8. bis 18. Jahrhundert. Berlin, S. 383–410.
  Susanne Haaf
  
• Differenzierungskriterien erbaulicher Textsorten des 17. Jahrhunderts. Operationalisierung von Ansätzen zu deren Beschreibung im Rahmen einer korpusbasierten Untersuchung. Vortrag im Rahmen des linguistischen Kolloquiums an der Universität Paderborn, 12. Juni 2019.
  Susanne Haaf
  
• Exploring TEI structures to find distinctive features of text types. Vortrag im Rahmen der TEI Conference. Graz, 16.–20. September 2019.
  Susanne Haaf
  
• Mehrdimensionale Beschreibung erbaulicher Textsorten des 17. Jahrhunderts. Jahrbuch für Germanistische Sprachgeschichte, 10(1), 324-344.
  Haaf, Susanne
  
• Merkmale registrieren oder textuelle Phänomene identifizieren? Zur Vereinbarkeit von automatischer und manueller Textsortenanalyse. Poster im Rahmen der DHd-Jahrestagung „Spielräume“, März 2020, Paderborn.
  Britt-Marie Schuster, Frauke Thielert, Susanne Haaf & Christopher Georg
  
• Sprachliche Muster – textliche Muster. Zu einem Wechselverhältnis und seiner Dynamik. Vortrag im Rahmen des Trier Center for Language and Communication - Lectures in Linguistics an der Universität Trier, 05. Februar 2020.
  Britt-Marie Schuster
  
• Alles eine Frage des (Textsorten)Stils? – Befunde zur Entwicklung von Pressetextsorten (17.-20 Jh.). Vortrag im Rahmen der Tagung „Die Geschichte der Textsorten“, Sarajevo, 18.–20. November 2021
  Britt-Marie Schuster
  
• Anleitung zur Verinnerlichung. Text, Bild und Textbild in erbaulichen Prosaschriften des 17. Jahrhunderts. Vortrag im Rahmen der Tagung „Text und Bild. Relationen und Funktionen in Texten vom 8. bis 18. Jahrhundert“, Universität Warschau, 17.– 19. Juni 2021 (virtuell).
  Susanne Haaf
  
• Das deutsche Textarchiv und die korpusbasierte Analyse komplexer Textmuster. Gastvorlesung im Rahmen der Vorlesungsreihe „Digital Humanities: Grundlagen, Methoden und Reflexion in interdisziplinärer Perspektive“, 2021.
  Kirsten, Linda & Boenig, Matthias
  
• Mixed Methods in Aktion. Das Projekt t.evo und die mehrdimensionale Annäherung an das Phänomen der Textsorte. Vortrag im Rahmen des Berliner DH-Kolloquiums an der BBAW (virtuell), 5. November 2021.
  Christopher Georgi, Susanne Haaf, Frauke Thielert & Linda Kirsten
  
• t.evo – Ein Tagset für die sprachpragmatische Annotation historischer Pressetexte (Guidelines).
  Frauke Thielert
  
• Textmuster erfassen – sprachliche Muster identifizieren. Überlegungen zu einem Wechselverhältnis am Beispiel des Kommentierens. Vortrag im Rahmen des 27. Deutschen Germanistentags an der Universität Paderborn, 25.-28. September 2022.
  Britt-Marie Schuster
  
• Exploiting the liaison of TEI text structuring and att.linguistic for the study of text types. Examples for 17th c. German devotional literature. Vortrag im Rahmen des ersten virtuellen Treffens der TEI SIG for Linguists, 18. März 2023.
  Susanne Haaf