Luddy School of Informatics, Computing, and Engineering, Indiana University, Bloomington, IN 47408.
Proc Natl Acad Sci U S A. 2022 Apr 26;119(17):e2117488119. doi: 10.1073/pnas.2117488119. Epub 2022 Apr 18.
It is a matter of debate whether a shrinking proportion of scholarly literature is getting most of the citations over time. It is also less well understood how a narrowing use of literature would affect the circulation of ideas in the sciences. Here, I show that the utilization of scientific literature follows dual tendencies over time: while a larger proportion of literature is cited at least a few times, citations are also concentrated more at the top of the citation distribution. Parallel to the latter trend, a paper’s future importance increasingly depends on its past citation performance. A random network model shows that the citation concentration is directly related to the greater stability of citation performance. The presented evidence suggests that the growing heterogeneity of citation impact restricts the mobility of research articles that do not gain attention early on. While concentration grows from the beginning of the studied period in 1970, citation dispersion manifests itself significantly only from the mid-1990s, when the popularity of freshly published papers also increased. Most likely, advanced information technologies to disseminate papers are behind both of these latter trends.
学术文献的引用是否呈现出随着时间的推移，越来越集中于少数文献的趋势，这是一个有争议的问题。文献使用范围的收窄将如何影响科学界思想的传播，这一点也不太为人理解。在这里，我表明，随着时间的推移，科学文献的利用呈现出两种趋势：尽管越来越多的文献至少被引用了几次，但引文也更加集中在引文分布的顶端。与后一种趋势平行的是，一篇论文的未来重要性越来越取决于其过去的引文表现。一个随机网络模型表明，引文的集中程度与引文表现的更大稳定性直接相关。所呈现的证据表明，引文影响力的日益增强的异质性限制了那些早期没有获得关注的研究论文的流动性。虽然集中趋势从 1970 年开始研究的初期就已经出现，但引文分散现象直到 20 世纪 90 年代中期才明显表现出来，当时新发表的论文的受欢迎程度也有所提高。很可能，先进的信息传播技术是这两个后期趋势背后的原因。
