Chairness investigates how the concept of a chair emerges as a learned and abstract category through computational perception. Drawing from both canonical designer chairs and synthetically generated variations, the project constructs a dataset that captures formal diversity while probing shared structural cues. A diffusion model trained on this dataset produces new chair forms that often depart from functional realism yet remain immediately recognizable. The results suggest that “chairness” is not tied to a single form or use, but to a set of internalized features shaping both human and machine recognition.