mlsanityData to 14 Oct 2026

MET: Theory-Grounded and Culture-Aware Multilingual Moral Reasoning

arXiv 2607.11736 ↗First posted 13 Jul 2026cs.CLcited by 0

On the map: LLM reasoning → Reasoning benchmarks and evaluation

Abstract (arXiv)

Language models are increasingly used for moral decision-making across diverse linguistic and cultural contexts, yet existing work overlooks multilinguality on three aspects: 1) multilingual evaluation benchmarks use direct translation, failing to adapt culture-specific items; 2) inference-time methods for moral reasoning rely on static, English-centric scaffolds and lack grounding in moral theory; 3) training methods for moral decision-making typically require expensive supervision from stronger models or human annotators. We address these gaps with three contributions. First, we introduce MCLASH, a multilingual moral decision-making benchmark to capture culturally situated moral intuitions and social norms across languages. Second, we propose MET (Multilingual Ethics with Theory-grounded reasoning), a two-step prompting method built on expert-curated, theory-based grounds drawn from psychology and philosophy: the model first selects situation- and culture-specific grounds, then reasons over them in the native language of the user. Third, we introduce MET-D (MET-Distillation), which enhances the second step through a self-distillation training stage that requires no external supervision. MET-D improves macro-F1 over the base model on all three models of different sizes and families (Qwen3-4B, Qwen3-8B, Gemma3-4B), by an average of 3.71 points on MCLASH and 4.23 on MMoralExceptQA, with a peak MCLASH gain of 12.94 points for Malay on Qwen3-8B. We further reveal that MET-D increases native-language reasoning by 62.13 points on average, and that beneficial grounds differ systematically across cultures. Together, these contributions open the path for culture-aligned, theory-grounded multilingual moral reasoning.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 43 of its 92 references that are in the map. First posted on arXiv 13 Jul 2026.

Builds on

43 of 92 references.

Closest ancestors · 15

Same direction, newest first

  1. 2025MoralReason: Generalizable Moral Decision Alignment For LLM Agents Using Reasoning-Level Reinforcement Learning15 Nov
  2. 2025MoReBench: Evaluating Procedural and Pluralistic Moral Reasoning in Language Models, More than Outcomes18 Oct
  3. 2025One Model, Many Morals: Uncovering Cross-Linguistic Misalignments in Computational Moral Reasoning25 Sep
  4. 2025Learn Globally, Speak Locally: Bridging the Gaps in Multilingual Reasoning7 Jul
  5. 2025Structured Moral Reasoning in Language Models: A Value-Grounded Evaluation Framework17 Jun
  6. 2025Are Language Models Consequentialist or Deontological Moral Reasoners?27 May
  7. 2025The Staircase of Ethics: Probing LLM Value Priorities through Multi-Step Induction to Complex Moral Dilemmas23 May
  8. 2025Language Matters: How Do Multilingual Input and Reasoning Paths Affect Large Reasoning Models?23 May
  9. 2025When Less Language is More: Language-Reasoning Disentanglement Makes LLMs Better Multilingual Reasoners21 May
  10. 2025Qwen3 Technical Report14 May
  11. 2025Moral Reasoning Across Languages: The Critical Role of Low-Resource Languages in LLMs28 Apr
  12. 2025Are Rules Meant to be Broken? Understanding Multilingual Moral Reasoning as a Computational Pipeline with UniMoral19 Feb
3 more
  1. 2024Ethical Reasoning and Moral Value Alignment of LLMs Depend on the Language we Prompt them in29 Apr
  2. 2024LangBridge: Multilingual Reasoning Without Multilingual Supervision19 Jan
  3. 2024MAPO: Advancing Multilingual Reasoning through Multilingual Alignment-as-Preference Optimization12 Jan

Other references · 28

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  3. 2024GPT-4o System Cardcited by 4,588
  4. 2022Large Language Models Encode Clinical Knowledgecited by 1,830
  5. 2020Aligning AI With Shared Human Valuescited by 723
  6. 2018Explaining Explanations: An Overview of Interpretability of Machine Learningcited by 508
  7. 2019Massively Multilingual Neural Machine Translation in the Wild: Findings and Challengescited by 257
  8. 2023Evaluating the Moral Beliefs Encoded in LLMscited by 215
  9. 2023Do All Languages Cost the Same? Tokenization in the Era of Commercial Language Modelscited by 192
  10. 2024BLEnD: A Benchmark for LLMs on Everyday Knowledge in Diverse Cultures and Languagescited by 173
  11. 2023Having Beer after Prayer? Measuring Cultural Bias in Large Language Modelscited by 153
  12. 2024Investigating Cultural Alignment of Large Language Modelscited by 143
16 more
  1. 2021Lawformer: A Pre-trained Language Model for Chinese Legal Long Documentscited by 116
  2. 2022When to Make Exceptions: Exploring Language Models as Accounts of Human Moral Judgmentcited by 93
  3. 2024DailyDilemmas: Revealing Value Preferences of LLMs with Quandaries of Daily Lifecited by 59
  4. 2023A Brief Report on LawGPT 1.0: A Virtual Legal Assistant Based on GPT-3cited by 46
  5. 2024Language Model Alignment in Multilingual Trolley Problemscited by 46
  6. 2023MoCa: Measuring Human-Language Model Alignment on Causal and Moral Judgment Taskscited by 44
  7. 2024Moral Alignment for LLM Agentscited by 39
  8. 2024Breaking the Curse of Multilinguality with Cross-lingual Expert Language Modelscited by 38
  9. 2024Self-Pluralising Culture Alignment for Large Language Modelscited by 31
  10. 2024JMMMU: A Japanese Massive Multi-discipline Multimodal Understanding Benchmark for Culture-aware Evaluationcited by 19
  11. 2024CIVICS: Building a Dataset for Examining Culturally-Informed Values in Large Language Modelscited by 16
  12. 2025CulFiT: A Fine-grained Cultural-aware LLM Training Paradigm via Multilingual Critique Data Synthesiscited by 15
  13. 2024Language Models are Alignable Decision-Makers: Dataset and Application to the Medical Triage Domaincited by 14
  14. 2025DRISHTIKON: A Multimodal Multilingual Benchmark for Testing Language Models' Understanding on Indian Culturecited by 13
  15. 2025CLASH: Evaluating Language Models on Judging High-Stakes Dilemmas from Multiple Perspectivescited by 10
  16. 2018The What, the Why, and the How of Artificial Explanations in Automated Decision-Makingcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024Whose Morality Do They Speak? Unraveling Cultural Bias in Multilingual Language Models
  2. 2025Cultural Bias in Large Language Models: Evaluating AI Agents through Moral Questionnaires
  3. 2026Untangling Input Language from Reasoning Language: A Diagnostic Framework for Cross-Lingual Moral Alignment in LLMs
  4. 2024Decoding Multilingual Moral Preferences: Unveiling LLM's Biases Through the Moral Machine Experiment
  5. 2025From Training to Generalization: Improving Moral Reasoning Through Pragmatic Inference
  6. 2025BengaliMoralBench: A Benchmark for Auditing Moral Reasoning in Large Language Models within Bengali Language and Culture
  7. 2024Do Moral Judgment and Reasoning Capability of LLMs Change with Language? A Study using the Multilingual Defining Issues Test
  8. 2025Diagnosing Moral Reasoning Acquisition in Language Models: Pragmatics and Generalization