The Berlin wall divided the city for four decades before it was torn down thirty years ago. However, in this open-access publication in Environment and Planning A (pdf), Milad Abbasiharofteh and I show that it is still shaping the evolution of knowledge networks. Using data on co-patents, co-publication and joint research grants, we model the evolution of the intra-city biotechnology knowledge network over 20 years. Separable Temporal Exponential Random Graph models (STERGM) allow us to study its underlying dynamics. In addition to the usual suspects, e.g. different types of proximities, we identify relations between organisations co-located in East or West Berlin being more likely to emerge than those bridging the former border. We explain this with the existence of distinct social foci.Accordingly, the paper’s contribution goes beyond showing the (still existing) relevance of the former Berlin wall. It highlights that history and place-specifics matter for the evolution of knowledge networks and that these are not captured by the usual proximity dimensions.