CSH researcher Niklas Reisz will present a talk within the seminar “Analyses of complex systems” on May 21, 2021 from 3–4 pm (CET).
If you want to join the talk, please send an e-mail to email@example.com
The global scientific output in terms of numbers of publications has been increasing rapidly—exponentially—in recent years, replacing and outdating historical knowledge at a tremendous rate. Only a few scientific milestones of exceptional quality have remained relevant over decades and consistently attract citations. From the perspective of scientometrics and science funding, this raises important questions. For how long is scientific work relevant? How long will it take before today’s work is forgotten? And how are milestone papers remembered differently?
To answer these questions, we study the temporal citation network of Physics works published in American Physical Society journals. We present a model to quantify the probability of attracting citations for individual publications based on their age and the number of citations they received in the past. We identify the bulk of publications that is forgotten in a largely homogeneous way as well as exceptional works that remain relevant for long periods of times.
We capture the forgetting as well as the tendency to cite popular works in a microscopic generative model of scientific citation networks. The model allows us to understand how some papers are cited while others are forgotten in the context of a growing citation network. Moreover, it allows us to estimate an expected citation landscape of the future, predicting that 95% of papers relevant in 2050 have not yet been written. We identify some conditions under which individual publications manage to achieve immortality.