A unique collaboration between physicists, data scientists and art historians has provided a fresh look at 500 years of art history.
The international team statistically analysed nearly 15,000 Western landscape paintings in an attempt to quantify their design principles. This revealed not only that composition patterns have evolved through history, but also that they characterise individual artists and artistic styles.
“Understanding how artistic expressions and design principles have changed over time is a central question in art history, aesthetics, and cultural evolution,” the researchers explain in a paper in Proceedings of the National Academy of Sciences.
Art has long functioned as a significant channel for human creativity and communication. Throughout history, it has evolved in a complex interplay with the social, technological and scientific environments of the time.
Studying art history, according to lead researcher Byunghwee Lee from the Korea Advanced Institute of Science and Technology, therefore represents “an effort to understand the creative process of humans as one of the essential natures of human beings, which needs to be understood to understand ourselves”.
Traditionally, the vast majority of art history research has employed qualitative methods, which tend to be interpretive and exploratory. The field has only recently been approached using numeric methods – made possible by the development of large-scale digital databases.
This new study analysed digital scans of 14,912 paintings, covering the Western renaissance to contemporary art, using an image dissection methodology, informed by information theory, to divide paintings into vertical and horizontal segments from the most to least prominent compositional features.
Horizontal segments included elements such as the sky, earth and atmospheric colour changes, while vertical segments included trees, plants, cliffs and buildings.
Interestingly, the study found that the positioning of these segments evolved.
Horizons, for example, have migrated upwards over the last few centuries. The skies in 17th century Baroque art frequently dominated the landscape, but during the Rococo and Romantic periods, the horizon moved up to around the midline of paintings. By the time the Realism and Impression periods rolled around, the horizon was positioned primarily in the upper third of the canvas.
Dissecting paintings in this quantitative way, the researchers explain, can “capture the unique compositional characteristics and systematic evolution of individual artist bodies of work, creation date time spans, and conventional style periods”.
But there is still plenty of scope to build on the study.
“Although the dataset used in this study includes some Japanese and Chinese landscape paintings, our dataset mainly focuses on paintings by European artists,” the authors acknowledge – creating a bias both in terms of gender and geography.
Their methods can act as a starting point to investigate the principles of artistic composition over a broader range of cultures and regions.
This approach could also be applied to other art forms – such as photography, film, typography and architecture – to reveal patterns not readily discernible to the individual eye.
Applying scientific knowledge to art or aesthetics may be viewed by many as reductionist – though scientists have argued that such studies are “efforts to understand the beauty of painting art in their own languages”.
According to Eleanor Gates-Stuart, artist and Professor of Creative Industries at Australia’s Charles Sturt University, there are also many aspects of the painting process – including emotive, gestural-driven actions by the artist – that cannot be extracted by scientific methods.
But she says this new algorithmic analysis “reveals a knowledge structure and methodology that is indeed a valuable systematic model”.
“Using statistical methods is certainly very useful expansion of integrating cross-disciplinary research, as shown here in this paper, especially in such a myriad of art history and aesthetics,” she adds.
Related reading: Art meets science