Defining Argumentative Discourse Units as Clauses: Psycholinguistic Evidence

Research output: Chapter in Book/Report/Conference proceedingChapter

5 Downloads (Pure)

Abstract

Identifying the smallest units of human argumentation remains a key challenge for computational models of argument. This study tested the assumption that argumentative discourse units (ADUs) can be best described as clauses. Two online experiments investigated the role of ADUs in human language processing(Experiment 1) and recall (Experiment 2), providing evidence that discourse comprehension might be influenced by syntactic depth. Experiment 1 analysed participants’ cued recall of pairs of clauses to identify whether they relied on clausal units when encoding information. Experiment 2 tested effects of manipulating the syntactic complexity (natural language – varying complexity, elementary language– one clause per sentence, or atomic language – one sub-clausal unit of information per sentence) on participants’ free recall of short encyclopaedic entries adapted from Wikipedia. Both experiments found small-to-medium effects suggesting that defining ADUs as clauses might be justified to a degree, with potential implications for computational models of human argumentation
Original languageEnglish
Title of host publicationProceedings of COMMA 2024
EditorsChris Reed, Matthias Thimm, Tjitze Rienstra
PublisherIOS Press BV
Pages277-288
Number of pages12
ISBN (Electronic)9781643685342
ISBN (Print)9781643685342
DOIs
Publication statusPublished - 27 Aug 2024

Publication series

NameComputational Models of Argument
ISSN (Print)09226389
ISSN (Electronic)18798314

Fingerprint

Dive into the research topics of 'Defining Argumentative Discourse Units as Clauses: Psycholinguistic Evidence'. Together they form a unique fingerprint.

Cite this