Page - (000188) - in Knowledge and Networks

182 the fact that the periphery (cluster 5) suddenly became the bridging core. Figure 9.3 shows that in 2009 cluster 5 was connected to all the other clusters but it had no loop. That is, its members were disconnected from each other, but they have become key structural brokers within the network. This result alone calls into question the very concept of network periphery. In addition, in 2009 a new bridging core emerged from the inconsistencies in cluster 4. The new bridging core indicates that inconsis- tencies were becoming structured and that a new collaboration pattern was being established from below. From 2010 onward, the development trajectory of IMAST remarkably phased away from a core–periphery model, and the pattern identified in 2009 gradually evolved into a new trajectory. In 2010 cluster 6 represented the main bridging core; like cluster 5 in 2009, it also had no loop. This evidence further confirms that struc- tural transformations were taking place at the periphery of the network. Moreover, in 2010 cluster 4 remained a bridging core connected through inconsistencies, just as it was in 2009. This process of transition from a classic core–periphery model culiminated in 2011, when the periphery suddenly disappeared. From 2011 to 2013, all clusters had loops, indicating that intercluster collaborations were occurring. The hypothesis that a core–periphery structure persisted from 2006 to 2013 can therefore be rejected. Instead, the collaboration network evolved toward a new model character- ized by multiple bridging cores and an increasing number of sporadic ties linking Fig. 9.3 Reduced graphs of IMAST collaboration networks over time. Nodes represent clusters of structurally equivalent actors; links represent intergroup relations; loops signal within-group con- nections. Red ties signal positive ties in blocks specified as null (bridging ties) (Source: Authors’ elaborations based on R&D collaboration within the technological district) L. Prota et al.