The critical investigation of paintings is essentially comparative. This limits the number of works of art that can be investigated by an art expert within a reasonable time. The recent availability of large digitized art collections allows for a scaling up of this analysis through the use of computer methods. Our research shows that simple physics- inspired measurements that are estimated from local spatial order models in paintings encode crucial information about the artwork. We present numerical scales that correspond well to canonical concepts of art history and reveal a historical and measurable evolutionary trend in the visual arts. They also allow us to distinguish between different artistic styles and works of art depending on the degree of local order in the paintings.
