# On the Visualisation of Argumentation Graphs to Support Text Interpretation

**Authors:** Hanadi Mardah, Oskar Wysocki, Markel Vigo, Andre Freitas  
**arXiv:** [2303.03235](https://arxiv.org/abs/2303.03235)  
**Date:** 2023-03-06  
**Summary page:** https://www.summarizepaper.com/en/paper/2303.03235v1/

## Key points

- Recent advancements in Natural Language Processing (NLP) methods, particularly in argumentation mining, have the potential to revolutionize text interaction and analysis of complex discourse and debates.
- A study investigated whether a graphic visualization of complex argumentation could enable a more critical interpretation of arguments.
- The study presented 10 argument topics in two models: PTM (text-based model) and a 2D AGM (structured representation of an argument).
- The structured modality (2D AGM) significantly outperformed the text-based one (PTM) in terms of attractiveness, perspicuity, efficiency, dependability, and stimulation.
- Participants found it easier to interpret arguments using AGs than textual representations.
- Participants chose more critical questions when they were less familiar with the topic being discussed.
- There is a trade-off between time spent interpreting an argument and critical interpretation; while introducing a temporal overhead, the additional structure forces users to be more critical and systematic when judging the quality of an argument.
- Both models had similar effects on working memory.
- The study's limitations include its small sample size (25 participants), which may not be representative enough to generalize its findings across different populations or tasks.
- Future studies are needed to understand how humans critically interpret complex arguments or debates better.
- Visualizing complex argumentation through 2D AGMs can support critical interpretations of arguments; especially for unfamiliar topics.
- The study's findings have implications for designing more effective tools and methods to support critical interpretation processes in various domains including education journalism and law.
- Future studies should aim to validate these results and examine their influence on different populations tasks and settings.

## Summary

The recent advancements in Natural Language Processing (NLP) methods, particularly in the field of argumentation mining, have the potential to revolutionize the way we interact with text and support the interpretation and analysis of complex discourse and debates. This study aimed to investigate whether a graphic visualization of complex argumentation could enable a more critical interpretation of arguments. The study presented 10 argument topics in two models: PTM (text-based model) and a 2D AGM (structured representation of an argument). The evaluation process for these models was based on task-related constructs such as time spent, working memory effort, workload, and user experience. The results showed that the structured modality (2D AGM) significantly outperformed the text-based one (PTM) in terms of attractiveness, perspicuity, efficiency, dependability, and stimulation. The participants also found it easier to interpret arguments using AGs than textual representations. Interestingly, participants chose more critical questions (CQs) when they were less familiar with the topic being discussed. This finding suggests that AG schemes can be supportive of the interpretation process for unfamiliar topics. The study also confirmed that there is a trade-off between time spent interpreting an argument and critical interpretation; while introducing a temporal overhead, the additional structure forces users to be more critical and systematic when judging the quality of an argument. Furthermore, both models had similar effects on working memory. The study's limitations include its small sample size (25 participants), which may not be representative enough to generalize its findings across different populations or tasks. Additionally, while this study provides empirical evidence supporting AGs' effectiveness in supporting critical interpretation processes for complex arguments or debates compared to text-based models like PTMs; future studies are needed to understand how humans critically interpret complex arguments or debates better. In conclusion, this study demonstrated that visualizing complex argumentation through 2D AGMs can support critical interpretations of arguments; especially for unfamiliar topics. The study's findings have implications for designing more effective tools and methods to support critical interpretation processes in various domains including education journalism and law. Future studies should aim to validate these results and examine their influence on different populations tasks and settings.