Publikation:

Towards an In-Depth Automated Framing Detection : Synergizing NLP Techniques and Formal Pragmatic Models

Lade...
Vorschaubild

Dateien

Yu_2-191ihua3855gn9.pdf
Yu_2-191ihua3855gn9.pdfGröße: 5.23 MBDownloads: 48

Datum

2024

Autor:innen

Herausgeber:innen

Kontakt

ISSN der Zeitschrift

Electronic ISSN

ISBN

Bibliografische Daten

Verlag

Schriftenreihe

Auflagebezeichnung

DOI (zitierfähiger Link)
ArXiv-ID

Internationale Patentnummer

Angaben zur Forschungsförderung

Projekt

Open Access-Veröffentlichung
Open Access Green
Core Facility der Universität Konstanz

Gesperrt bis

Titel in einer weiteren Sprache

Publikationstyp
Dissertation
Publikationsstatus
Published

Erschienen in

Zusammenfassung

In NLP, recent years have witnessed a growing interest in automatically detecting framing from large-scale text collections. However, most existing studies focus heavily on framing effects arising from topic selection. In contrast, framing effects evoked by subtle rhetorical effects of topic-agnostic pragmatic cues remain understudied. This trend oversimplifies framing as a mere matter of topic coverage, providing only a limited understanding of the intricate mechanics of framing: as pointed out by earlier social science studies on framing theory, framing not only concerns what information is presented, but also how information is presented so that the perception of the audiences is potentially biased. Hence, subtle language choice constitutes a crucial aspect of framing.

However, little is known about the linguistic properties of framing: we are not aware of any research in the social sciences or NLP that provides a theoretical modelling of framing which accounts for the effects of individual linguistic components. To address this gap, our research aims to explore linguistic cues that contribute to framing effects. We seek to develop theoretical models and computational tools that not only incorporate framing induced by topic selection, but also framing arising from subtle linguistic manipulations.

This work commences with a comprehensive survey of works on framing in social science and NLP. Based on the survey, we criticize that the prevailing topic-centric methodology in NLP has especially neglected a crucial aspect of framing: the interplay between the information disseminated by the media and the cognitive processes of the individuals. By connecting this neglected aspect to pragmatic theories, we propose a formal model of framing that models the media-individual interplay. We further apply this model to elucidate the significance of various topic-agnostic pragmatic cues in the media-individual interplay. We substantiate the significant role of these deep pragmatic cues with three large-scale experiments using a German-language dataset with over 8 million tokens, bringing novel insights into the linguistic properties of framing. Especially, our work includes experiments on effects evoked by two phenomena that are well-studied in theoretical linguistics but remain understudied with regard to framing: presuppositions and modal particles. We demonstrate that automated framing detection can benefit from the conscious integration of theoretically motivated pragmatic features as such.

Overall, this work features a step towards successfully incorporating theoretical linguistic knowledge into NLP applications. It makes the following contributions:

Theoretical Contributions: We propose the first formal model of framing that captures the effects of individual linguistic cues. Based on this model, we illustrate the crucial role of a variety of topic-agnostic pragmatic cues in framing, which have not been studied by earlier work. We put forward the notion of rhetorical framing to characterize the framing effect of such cues, a new dimension of framing besides the well-studied topical framing.

Methodological Contributions: We release two open-source tools to facilitate future work on linguistically informed framing detection: the statistically compiled lexical resource RMFV (Refugees and Migration Framing Vocabulary) for the analyses of topical framing within the discourse of immigration, and the web App RheFrame (Rhetorical Framing Explorator) for the automated annotation of a wide range of rhetorical framing cues. With a series of large-scale experiments applying these tools, we show the influential role of rhetorical framing cues in uncovering framing strategies at fine-grained linguistic levels.

Zusammenfassung in einer weiteren Sprache

Fachgebiet (DDC)
400 Sprachwissenschaft, Linguistik

Schlagwörter

Natural Language Processing, Machine Learning, Large Language Models, Artificial Intelligence

Konferenz

Rezension
undefined / . - undefined, undefined

Forschungsvorhaben

Organisationseinheiten

Zeitschriftenheft

Zugehörige Datensätze in KOPS

Zitieren

ISO 690YU, Qi, 2024. Towards an In-Depth Automated Framing Detection : Synergizing NLP Techniques and Formal Pragmatic Models [Dissertation]. Konstanz: Universität Konstanz
BibTex
@phdthesis{Yu2024Towar-71516,
  year={2024},
  title={Towards an In-Depth Automated Framing Detection : Synergizing NLP Techniques and Formal Pragmatic Models},
  author={Yu, Qi},
  address={Konstanz},
  school={Universität Konstanz}
}
RDF
<rdf:RDF
    xmlns:dcterms="http://purl.org/dc/terms/"
    xmlns:dc="http://purl.org/dc/elements/1.1/"
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:bibo="http://purl.org/ontology/bibo/"
    xmlns:dspace="http://digital-repositories.org/ontologies/dspace/0.1.0#"
    xmlns:foaf="http://xmlns.com/foaf/0.1/"
    xmlns:void="http://rdfs.org/ns/void#"
    xmlns:xsd="http://www.w3.org/2001/XMLSchema#" > 
  <rdf:Description rdf:about="https://kops.uni-konstanz.de/server/rdf/resource/123456789/71516">
    <dcterms:available rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2024-12-02T10:15:58Z</dcterms:available>
    <dc:contributor>Yu, Qi</dc:contributor>
    <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/>
    <dcterms:issued>2024</dcterms:issued>
    <dc:rights>terms-of-use</dc:rights>
    <dspace:isPartOfCollection rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/45"/>
    <dcterms:title>Towards an In-Depth Automated Framing Detection : Synergizing NLP Techniques and Formal Pragmatic Models</dcterms:title>
    <dcterms:rights rdf:resource="https://rightsstatements.org/page/InC/1.0/"/>
    <dcterms:hasPart rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/71516/4/Yu_2-191ihua3855gn9.pdf"/>
    <dcterms:abstract>In NLP, recent years have witnessed a growing interest in automatically detecting framing from large-scale text collections. However, most existing studies focus heavily on framing effects arising from topic selection. In contrast, framing effects evoked by subtle rhetorical effects of topic-agnostic pragmatic cues remain understudied. This trend oversimplifies framing as a mere matter of topic coverage, providing only a limited understanding of the intricate mechanics of framing: as pointed out by earlier social science studies on framing theory, framing not only concerns what information is presented, but also how information is presented so that the perception of the audiences is potentially biased. Hence, subtle language choice constitutes a crucial aspect of framing. 

However, little is known about the linguistic properties of framing: we are not aware of any research in the social sciences or NLP that provides a theoretical modelling of framing which accounts for the effects of individual linguistic components. To address this gap, our research aims to explore linguistic cues that contribute to framing effects. We seek to develop theoretical models and computational tools that not only incorporate framing induced by topic selection, but also framing arising from subtle linguistic manipulations. 

This work commences with a comprehensive survey of works on framing in social science and NLP. Based on the survey, we criticize that the prevailing topic-centric methodology in NLP has especially neglected a crucial aspect of framing: the interplay between the information disseminated by the media and the cognitive processes of the individuals. By connecting this neglected aspect to pragmatic theories, we propose a formal model of framing that models the media-individual interplay. We further apply this model to elucidate the significance of various topic-agnostic pragmatic cues in the media-individual interplay. We substantiate the significant role of these deep pragmatic cues with three large-scale experiments using a German-language dataset with over 8 million tokens, bringing novel insights into the linguistic properties of framing. Especially, our work includes experiments on effects evoked by two phenomena that are well-studied in theoretical linguistics but remain understudied with regard to framing: presuppositions and modal particles. We demonstrate that automated framing detection can benefit from the conscious integration of theoretically motivated pragmatic features as such. 

Overall, this work features a step towards successfully incorporating theoretical linguistic knowledge into NLP applications. It makes the following contributions:

Theoretical Contributions: We propose the first formal model of framing that captures the effects of individual linguistic cues. Based on this model, we illustrate the crucial role of a variety of topic-agnostic pragmatic cues in framing, which have not been studied by earlier work. We put forward the notion of rhetorical framing to characterize the framing effect of such cues, a new dimension of framing besides the well-studied topical framing. 

Methodological Contributions: We release two open-source tools to facilitate future work on linguistically informed framing detection: the statistically compiled lexical resource RMFV (Refugees and Migration Framing Vocabulary) for the analyses of topical framing within the discourse of immigration, and the web App RheFrame (Rhetorical Framing Explorator) for the automated annotation of a wide range of rhetorical framing cues. With a series of large-scale experiments applying these tools, we show the influential role of rhetorical framing cues in uncovering framing strategies at fine-grained linguistic levels.</dcterms:abstract>
    <dspace:hasBitstream rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/71516/4/Yu_2-191ihua3855gn9.pdf"/>
    <dc:language>eng</dc:language>
    <foaf:homepage rdf:resource="http://localhost:8080/"/>
    <dc:creator>Yu, Qi</dc:creator>
    <dc:date rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2024-12-02T10:15:58Z</dc:date>
    <bibo:uri rdf:resource="https://kops.uni-konstanz.de/handle/123456789/71516"/>
    <dcterms:isPartOf rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/45"/>
  </rdf:Description>
</rdf:RDF>

Interner Vermerk

xmlui.Submission.submit.DescribeStep.inputForms.label.kops_note_fromSubmitter

Kontakt
URL der Originalveröffentl.

Prüfdatum der URL

Prüfungsdatum der Dissertation

October 11, 2024
Hochschulschriftenvermerk
Konstanz, Univ., Diss., 2024
Finanzierungsart

Kommentar zur Publikation

Allianzlizenz
Corresponding Authors der Uni Konstanz vorhanden
Internationale Co-Autor:innen
Universitätsbibliographie
Ja
Begutachtet
Diese Publikation teilen