When we're making a study of a network or formulating an explanation for outcomes produced by a network, we need things to examine. We are prone to make a "thingy" of nothing or of an ongoing processing. Alfred North Whitehead coined the term "misplaced concreteness" to separate his "process philosophy" from all the scientific studies that had made things of processes in order to study them objectively.
In the Prince of Networks, Graham Harman characterized Whitehead as the grandfather of Bruno Latour's Actor Network Theory. Latour takes exception to scientific explanations that add a false dimension to their descriptions of networks. He expects explanations to emerge from sufficient complexity of descriptions, rather than conforming the data to established explanations. He treasures Whitehead's investigative principle of "beginning and ending with wonder".
As I've pondered how all this might support our rethinking the concreteness of nodes in a network, I've made a lot of associations to the possibility of nodeless networks. If there's nothing there where connections come together, we have to wonder what the intersection is about. The connections would meet with wonder and explore what it's about. There could be no "misplaced concreteness" in a absence of nodes.
What could occur at "connections over nothing" is significant learning. The unknowns could come to the forefront of the experience. The exploration of combined questions, possibilities, hypotheses, and contrasting descriptions could replace the so-called "learning" of information through network connections. The takeaways from the encounters could enrich each exploration and better articulate the intersection.
This is a reversal of preexisting nodes forming new connections. This proposes preexisting connections resulting in emergent nodes. The coming together is the constant process. The resulting nodes come and go. The process begins and ends with wonder.