Disparities In Negation Understanding Across Languages In Vision-Language Models

TL;DR

This paper introduces a multilingual negation benchmark to evaluate vision-language models across diverse languages, revealing disparities in negation understanding linked to linguistic features and model architecture.

Contribution

It presents the first human-verified multilingual negation benchmark covering seven languages, evaluating multiple VLMs and analyzing linguistic factors affecting negation comprehension.

Findings

Standard CLIP performs at or below chance on non-Latin scripts.

MultiCLIP achieves the most accurate and consistent negation understanding.

Negation correction methods improve performance variably across languages.

Abstract

Vision-language models (VLMs) exhibit affirmation bias: a systematic tendency to select positive captions ("X is present") even when the correct description contains negation ("no X"). While prior work has documented this failure mode in English and proposed solutions, negation manifests differently across languages through varying morphology, word order, and cliticization patterns, raising the question of whether these solutions serve all linguistic communities equitably. We introduce the first human-verified multilingual negation benchmark, spanning seven typologically diverse languages: English, Mandarin Chinese, Arabic, Greek, Russian, Tagalog, and Spanish. Evaluating three VLMs - CLIP, SigLIP, and MultiCLIP - we find that standard CLIP performs at or below chance on non-Latin-script languages, while MultiCLIP achieves the highest and most uniform accuracy. We also evaluate SpaceVLM, a proposed negation correction, and find that it produces substantial improvements for several languages - particularly English, Greek, Spanish, and Tagalog - while showing varied effectiveness across typologically different languages. This variation reveals that linguistic properties like morphology, script, and negation structure interact with model improvements in fairness-relevant ways. As VLMs are deployed globally, multilingual benchmarks are essential for understanding not just whether solutions work, but for whom.