I am developing computational methods for comparing the features and meanings of scientific image data over time.
These methods combine bibliometric analysis with different image featurization techniques to identify groups of images that facilitate the social evolution of scientific knowledge.
What is a Scientific Image?
A scientific image is any image made by a scientist.
Scientists make and use images—2D visual representations like illustrations, diagrams, graphs, maps, photographs—to record and communicate data, negotiate concepts, and produce knowledge.
Images are central of the process and progress of science.
Just like scientific knowledge changes over time, so do the features and meanings of scientific images. Historians and philosophers studying science used to think that scientific images were becoming more accurate over time.
For example, one could interpret scientific images representing the human brain over many centuries as changing to become more accurate as scientific knowledge about the brain became more accurate. But are knowledge and images of the brain becoming more accurate, or true to reality, over time?
During the mid-20th century, some historians and philosophers studying why scientific knowledge changed over time came to appreciate the importance of its social and cultural origins. Rather than being understood as shaped by “reality,” scientific knowledge was thought to be shaped by the cultures in which it was produced and by the problems it was designed to address. We didn't have to think about scientific knowledge as becoming more accurate or true, but rather as something that changes over time for different reasons.
Historians, philosophers, and sociologists of science wondered: if scientific images were not changing to represent a more accurate view of the world, then what do scientific images represent, and why do they change? In the 1980s, these scholars began to document how images were direct products of material and social experimental practices.
By their account, no one scientific image of the brain inherently represents reality more truthfully than another; rather, the features and meanings of scientific images of the brain are different, and these differences render specific images better or worse at addressing specific problems in specific contexts.
HOW SHOuld WE learn about SCIENTIFIC IMAGES?
Researchers working in many different disciplines are interested in learning about the features and meanings of scientific images. Biologists, art historians, philosophers, cultural anthropologists, and artificial intelligence researchers want to know why scientific images look the way they do, and mean what they mean. There are lots of methods you could apply to learn these things about scientific images.
When researchers try to explain why a scientific image looks the way it does, or means what it means, they often track the material and experimental practices that resulted in that image getting made. And that makes sense! An image will look a certain way and have a certain meaning based on the actions of the specific image maker, the technology they used to make it, and the physical entities involved in its generation. But this is not a complete account of how scientific images are shaped and get their meaning.
Historians, philosophers, meta-scientists, and scientists themselves rarely have access to the direct material and experimental practices by which scientific images are made, such that images must have representational capacities beyond the contexts of their direct production.
In my PhD work, I showed that scientists in fields like microbiology and cell biology studying the same physical entities appear to design images that more often share visual features with images they have cited (please contact me for a copy of my dissertation). The features and meanings of scientific images, then, are shaped by more than material and experimental practices. The features and meanings of scientific images are constrained by the features and meaning of other images.
To learn why a specific scientific image looks the way it does, and means what it means, we need to study other images.
Back to the brain example: scientific illustrations of the brain are certainly changing in response to material and experimental contexts, but the features and meanings of new images of the brain were likely constrained by the specific features and meanings of prior images of the brain to which anatomists, physicians, and medical illustrators had access at the time.
Co-citation (author) network for the “microbial biofilm” concept through 1974:
(Made with VoS Viewer)