More Than Meets the Eye: Measuring the Semiotic Gap in Vision-Language Models via Semantic Anchorage
For researchers in vision-language understanding, this work identifies a fundamental limitation of current VLMs in handling abstract meaning, highlighting the need for iconographic abstraction in visual inputs.
The paper introduces DIVA, a benchmark for evaluating vision-language models on idiomatic vs. literal interpretations of noun compounds, and proposes a metric (Δ) to measure the semantic alignment gap. Experiments on 8 VLMs reveal a consistent Literal Superiority Bias, where higher visual fidelity correlates with weaker symbolic alignment, indicating that model scale alone does not resolve this bias.
Vision-Language Models (VLMs) excel at photorealistic generation, yet often struggle to represent abstract meaning such as idiomatic interpretations of noun compounds. To study whether high visual fidelity interferes with idiomatic compositionality under visual abstraction, we introduce DIVA, a controlled benchmark that replaces high-fidelity visual detail with schematic iconicity by generating paired, sense-anchored visualizations for literal and idiomatic readings. We further propose Semantic Alignment Gap ($Δ$), an architecture-agnostic metric that quantifies divergence between literal and idiomatic visual grounding. We additionally introduce a directional signed bias $b(t)$ to separately measure the direction and strength of literal preference. Evaluating 8 recent VLMs, we reveal a consistent Literal Superiority Bias: model scale alone does not resolve literal preference, and increased visual fidelity is associated with weaker symbolic alignment, suggesting cognitive interference from hyper-realistic imagery. Our findings suggest that improving compositional understanding requires iconographic abstraction of visual input and anchoring interpretation and generation in intended meaning.