Network science did not begin when the phrase was coined.It began much earlier, when scholars in different fields learned to treat relations as objects of study in their own right.The path from Euler's Knigsberg bridges to the higher-order and learning-based models of the mid-2020s is therefore not a straight line but a convergence.Mathematics supplied abstraction.Sociology supplied empirical discipline.Physics supplied generative models, universality arguments, and an appetite for large systems.Computer science supplied data structures, algorithms, and, later, machine learning.Over time, these traditions ceased to run in parallel and started to speak to one another.This essay reconstructs that convergence.It moves from graph theory, electrical circuits, and sociometry to random graphs, citation networks, small-world structure, preferential attachment, community detection, motifs, epidemic processes, temporal and multilayer networks, higherorder representations, and graph learning.Throughout, the argument is simple.Network science became a field not when it found a single method, but when it recognized a common problem across domains: how structure constrains flow, diffusion, coordination, vulnerability, and discovery.By 2026, that problem is being studied on richer objects than the simple graph alone, yet the core intuition remains unchanged.In many of the most important systems we know, what matters is not only what the units are.It is how they are tied together, how those ties evolve, and what new phenomena appear because of that arrangement.
Joao Tiago Aparicio (Thu,) studied this question.