ZurichNLP/ArgMining-2026-UZH-Shared-Task

Reconstructing the Reasoning in United Nations Resolutions

This repository contains the dataset for the UZH Shared Task @ ArgMining Workshop 2026, co-located with ACL 2026.
The shared task focuses on recovering paragraph-level argumentative structure in highly formal, legal-political documents, specifically United Nations resolutions and recommendations.

The dataset supports two subtasks:

1. Argumentative paragraph classification
2. Argumentative relation prediction between paragraphs

---

Claimer

The content of this publication has not been approved by the United Nations and does not reflect the views of the United Nations or its officials or Member States. UN-RES should only be used for research purposes.

---

Contact

Please contact the shared task organizers at University of Zurich for questions.

Task Overview

United Nations resolutions encode collective reasoning at scale through carefully structured preambles and operative clauses.
This shared task evaluates how well systems can reconstruct this implicit reasoning structure.

Participants are expected to build systems that:

• Identify whether a paragraph is preambular or operative
• Assign one or more argumentative tags
• Predict argumentative relations between paragraphs

Only open-weight language models with ≤ 8B parameters are permitted.

---

Subtasks

Subtask 1: Argumentative Paragraph Classification

For each paragraph, systems must predict:

• Paragraph type: preambular or operative
• Argumentative tags: multi-label classification over a predefined tag set

Subtask 2: Argumentative Relation Prediction

For each paragraph, systems must:

• Identify related paragraphs (by index)
• Assign one or more relation types:
  • supporting
  • contradictive
  • complemental
  • modifying

---

Dataset Description

Languages

• English
• French

Granularity

• Paragraph level

Splits

Training Set (parsed_data_en)

• Source: UN-RES dataset Gao et al., 2025
• Size: 2,695 UN resolutions
• Language: French (with machine-generated English translations)
• Annotation: paragraph-level argumentative structure

Test Set (parsed_data_fr)

• Source: UNESCO International Conference on Education (1934–2008)
• Size: 45 parsed documents (each may contain up to three resolutions in JSON)
• Language: French
• Annotation: paragraph-level (held out for evaluation)
• Validation set: none

Tags

• See education_dimensions_updated.csv

---

Data Format

All data are provided in JSON format following a fixed schema.

Example (simplified)

"TEXT_ID": "ICPE-25-1962_RES1-FR_res_54",
  "RECOMMENDATION": 54,
  "TITLE": "LA PLANIFICATION DE L'ÉDUCATION",
  "METADATA": {
    "structure": {
      "doc_title": "ICPE-25-1962_RES1-FR",
      "nb_paras": 58,
      "preambular_para": [], 
      "operative_para": []
      "think": ""      
    }
  },
  "body": {
    "paras": [
      {
        "para_number": 1,
        "para": "La Conférence internationale de l'instruction publique, Convoquée à...",
        "type": null,      
        "tags": [],     
        "matched_paras": [],     
        "think": "",
        "para_en": "The International Conference on Education, convened in ..."
      },
      ...
    ]
  }
}