Decomposability in Knowledge Structures and Its Impact on the Usefulness of Inventions and Knowledge-base Malleability
利用1984至1994年全球半导体产业专利数据，研究组织知识库结构（从完全可分解到不可分解）如何影响发明有用性（以专利引用衡量）和知识库可塑性，发现近乎可分解结构效果最佳。
We use patent data from the worldwide semiconductor industry from 1984 to 1994 to study the effect of the structure of organizational knowledge bases, or the patterns of coupling between their elements of technical knowledge, on the usefulness of inventions and knowledge-base malleability. We argue that organizational variations in coupling patterns between knowledge elements can be reflected in a spectrum of knowledge-base structures—varying from fully decomposable (the knowledge base is composed of distinct clusters of knowledge elements coupled together with no significant ties between clusters) through nearly decomposable (knowledge clusters are discernable but are connected through cross-cluster couplings) to non-decomposable (no knowledge clusters emerge, as the couplings are pervasively distributed)—and that organizations may differ in the way they use their knowledge because of variations in their knowledge-base structure, rather than because of differences in the knowledge elements themselves. Results show that a nearly decomposable knowledge base increases the usefulness of the inventions generated from it, as measured by patent citations, and also the knowledge base's malleability or capacity for change.