Reasoning Gestures: LLM-Inferred Communicative Functions for Co-Speech Gesture Generation

When humans speak, gestures often realize communicative functions (e.g., emphasizing key content, indicating referents, or depicting concepts). Most co-speech gesture generation methods, however, condition primarily on low-level cues (speech audio or transcripts) and do not explicitly represent these higher-level functions, leading to motions that synchronize with speech but remain under-specified in communicative meaning. To address this gap, we introduce Reasoning Gestures, a reasoning-grounded framework that conditions generation on interpretable communicative-function descriptions. We curate InG: Interpreted Gestures by augmenting BEAT-2 with sentence-level VLM-inferred gesture-function annotations that describe the discourse role of each gesture segment in context. These labels are observer-centric interpretations produced via constrained VLM reasoning rather than claims about speakers’ private mental states. We further propose a Reasoning Gesture Motion Tokenizer that injects function signals into tokenized motion representations, enabling function-aware synthesis that is temporally aligned and communicatively grounded. Experiments on BEAT-2 show consistent gains in semantic consistency and overall quality, establishing new state-of-the-art.