TL;DR
This invited talk argues that traditional, logic-centric views of argumentation fall short of capturing the messy reality of human discourse. It proposes that computational argumentation needs to embrace the multifaceted nature of communication, including common sense, emotions, and social context, to effectively model and analyze how we argue.
What Problem Does This Paper Address?
The paper tackles the challenge of building computational systems that can understand and engage in argumentation. It highlights a significant disconnect between the idealized, logical models of argumentation often found in textbooks and the way humans actually argue in the real world. Human argumentation is not just about logical structures; it's a rich, complex process deeply intertwined with common sense, emotions, and social context. Current computational approaches, often rooted in formal logic, struggle to capture this complexity, leading to systems that are brittle and fail to reflect the nuances of human interaction. The talk aims to shed light on how to bridge this gap by exploring recent advancements in computational argumentation that attempt to address these more human-centric aspects.
Methods
Not discussed in the paper. This is an invited talk abstract, not a research paper detailing specific methods.
Datasets
Not discussed in the paper.
Key Insights and Ideas
- The "Real World" is Messy: The classical view of argumentation as purely logical structures is a poor reflection of how humans actually communicate. Real-world arguments are far more complex and less structured.
- Argumentation is Multifaceted: Human argumentation relies heavily on common sense, emotions, and social context, not just logical deduction. These elements are crucial for effective communication and persuasion.
- Humans are Good at Arguing: Despite the complexity, humans are remarkably adept at navigating these multifaceted aspects of argumentation. This suggests that computational systems need to learn from and emulate these human capabilities.
- Computational Argumentation's Long Road: Building systems that can truly understand and participate in argumentation is a significant and ongoing challenge. Current efforts are still in their early stages.
- Beyond Logic: The talk signals a shift towards computational approaches that move beyond purely logical analysis to incorporate aspects like argument quality, reasoning processes, fallacy detection, and the dynamics of online argumentation.
Key Results
Not discussed in the paper. This is an invited talk abstract.
Limitations and Future Work
The primary limitation acknowledged is the inherent difficulty and "long way to go" for computational argumentation to fully capture the complexities of human discourse. The talk itself is framed as an exploration of "several recent attempts" to address these challenges, suggesting that the field is actively working on solutions. Future work implicitly involves developing computational models that can effectively handle:
- Argument quality assessment.
- Understanding the underlying reasoning in arguments.
- Detecting and dealing with fallacies.
- Modeling the unique dynamics of online argumentation.
Summary updated