Understanding Function Shift of Representative Words based on the Analysis of "Attention" in Psychology

By taking representative words of “attention” as an example, we study the shifting word function with the time passing to detect the development status of attention-related research under the psychology background. Taking into consideration both rank/importance and information entropy/usage scene complexity, we characterize the functionality of words as four types of tree parts, each of which alone or associated with “attention” is a representation of different growth states within the scientific domain. 

From root words we can find concepts (e.g., selective-attention, crossmodal) laying the groundwork for future development; trunk-branch words (e.g., broadbent, socio-emotional) indicate research areas developing unpredictably; young-leaf words (e.g., manifold, two-process) suggests controversial topics hotly debated; dead-leaf words (e.g., non-hyperactive) demonstrate subjects becoming obsolete.

Our study expands traditional domain development analysis that usually only considered ascending/descending research trends to a multi-layered word function positioning analysis. From a hierarchical tree of words, we design a visual analysis framework that lays the foundation for a comprehensive and visual capture of a feature word’s knowledge structure and development overtime, providing us with a fresh perspective on understanding domain development status. In the future, changes in word functions over different time periods will be examined to facilitate a better understanding of domain development from a dynamic perspective.