Automated identification of discourse markers using the NLP approach - The case of okay

Oct 15, 2026

Upload by Franzy

This study evaluates how well a Natural Language Processing (NLP) approach can detect discourse markers (DMs), specifically the token okay, in transcribed spoken data. Human referees judged 138 concordance lines as DM or non-DM, then researchers applied a Python rule set based on POS tagging from NLTK and compared script output with human-annotated references. Results reached 92% accuracy, with remaining errors linked to insufficient punctuation details, implying NLP can detect DMs but requires proper punctuation.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Abstract

Introduction

Discourse Markers (DMs)

Background of the Study

| | | | | --- | --- | --- | | Australian Journal ofApplied Linguistics, 4 (3), 119–131 (2021) [https://doi.org/10.29140/ajal.v4n3.538](https://doi.org/10.29140/ajal.v4n3.538) | | | | Automated identification of discourse markers using the NLP approach: The case of okay | | | | ABDULAZIZ SANOSI a
MOHAMED ABDALLA b | | | | a Prince Sattam bin Abdulaziz University,
SAUDI ARABIA [a.assanosi@psau.edu.sa](a.assanosi@psau.edu.sa) | b Prince Sattam bin Abdulaziz University,
SAUDI ARABIA [m.abdalla@psau.edu.sa](m.abdalla@psau.edu.sa) | | | Abstract
This study aimed to examine the potentials of the NLP approach in detecting discourse markers (DMs), namely okay, in transcribed spoken data. One hundred thirty-eight concordance lines werepresented to human referees to judge the functions of okay in them as a DM or Non-DM. After that, the researchers used a Python script written according to the POS tagging scheme of the NLTK library to set rules for identifying cases where okay is used as non-DM. The output of the script was compared to the reference human-annotated data. The results showed that the script could accurately identify the function of okay as DM or non-DM in 92% of the cases. The inaccuracy of detecting therest was found to be caused by a lack of proper and detailed punctuations. The main implications of the results are that new NLP approaches can detect DMS; however, proper punctuation is required to enable the proper identification of DMs. In accordance with the findings, the researcher recommended adopting the approach after conducting further comprehensive studies.
Keywords: discourse markers, nlp, machine learning, corpus linguistics, discourse analysis | | | | Introduction
In linguistics, discourse is “an utterance whose magnitude is bigger than the sentence”(Jabeen, Rai,& Arif, 2011, p. 69) . It usually occurs due to “interaction between two or more participants who start talking to achieve a communicative goal” (Guo, 2015, p. 69) . As such an interaction entails that discourse should be coherent and logically presented, language users tend to employ various | | | Copyright: © 2021 Abdulaziz Sanosi & Mohamed Abdalla. This is an open access article distributed under the terms of the Creative Commons Attribution Non-Commercial 4. 0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within this paper. linguistic elements to maintain the coherence and relevance of their speech to make this interaction more effective. These elements have additional functions such as filling pauses, emphasising specific points, eliciting feedback, and signalling new topics. As the case with their various functions, many terms have been suggested referring to such elements, the most common of which is Discourse Markers (DMs) . Researchers investigate and analyse DM using different approaches to explore their function, structure, and distribution in the discourse. New methods of linguistic analysis have prospered recently, where the vast advances in technology have made using the computer for linguistic analysis a preferable trend. Corpus-based studies and computational linguistics are two disciplines that can stand as a representation of this breakthrough. It is now possible to analyse corpora of millions of words to investigate patterns of language use, frequency of occurrence, words collocation and colligation. When it comes to DMs, however, contextual difficulties that may harden the identification of DMs in the given discourse arise because DMs are extremely context-dependent (Furkó, 2020) . For this reason, determining DMs within corpora entails manual analysis, which may violate the core concept behind using a corpus-based approach, i.e., automatic analysis of vast bodies of text. The current study investigates such a problem and proposes dealing

FAQs

What goal does the study pursue regarding the discourse marker “okay”?

The study examines the potential of an NLP approach to detect “okay” as a discourse marker versus non-marker in transcribed spoken data.

How were “okay” cases identified and evaluated in the research?

Researchers used Python rules based on POS tagging (via NLTK) to flag instances where “okay” functions as non-DM, then compared the script output with human reference annotations.

What factor caused most of the remaining identification errors?

The inaccuracy for the non-detected cases was attributed to a lack of proper and detailed punctuations.

Oct 15, 2026

Upload by Franzy