The theoretical literature on technological changes distinguishes between paradigmatic changes and changes in trajectories. Recently several scholars have performed empirical studies on the way technological trajectories evolve in specific industries, often by predominantly looking at the artifacts. Much less - if any - empirical work has been done on paradigmatic changes, even though these have a much more profound impact on today's industry. It follows from the theory that such studies would need to focus more on the knowledge level than on the artifact level, raising questions on how to operationalize such phenomena. This study aims to fill this gap by applying network-based methodologies to knowledge networks, represented here by patents and patent citations. The rich technological history of telecommunications switches shows how engineers in the post-war period were confronted with huge challenges to meet drastically changing demands. This historical background is a starting point for an in-depth analysis of patents, in search of information about technological direction, technical bottlenecks, and engineering heuristics. We aim to identify when such changes took place over the seven different generations of technological advances this industry has seen. In this way we can easily recognize genuine paradigmatic changes compared to more regular changes in trajectory.