How do research fields evolve? Are there universal patterns behind the growth and decay of research areas? And can we better understand the roots of innovation?
These are the questions at the core of a recent article published by the Interaction Data Lab at the Learning Planet Institute. Using data from 1.5M articles from the preprint repository arXiv, they collected the publication history of 175 research fields from Physics, Maths, and Computer Science.
Some of these fields have peaked in the 1990's, with only little activity today (such as High Energy Physics), while other fields are booming (such as AI). Some have been of importance for decades, while other have mostly gained interest for a few years. Despite this diversity, researchers in the team found that all fields follow a well-defined rise and fall pattern, characterised by a peak time (when the field is at peak activity relative to the first article) and a width (for how long there is a sustained activity). By estimating these parameters for each field, the researchers could align all curves on a universal "law" of evolution, allowing them to compare these different fields on the same terms.
Using this method, the team then looked at the characteristics shared by scientists and articles at different stages of evolution of a research field: its creation, adoption, peak, early and late decay. They found that early stages, when the field is in its most "innovative" phase, are characterised by small interdisciplinary teams of early career researchers publishing disruptive work, while late phases exhibit the role of specialised, large teams building on the previous works in the field.
As such, these results unravel key characteristics shared by the "pioneers" in research: young, interdisciplinary, risk taking. Future work will extend these findings to larger databases going beyond the technical fields studied.
For more information, you can read the full paper here:
and read this Twitter thread: