Foundations of Network Science: History, Concepts, and Paradigms

The study of networks provides a rigorous mathematical and conceptual framework for quantifying and analyzing the patterns of connections that define systems ranging from social collectives to biological organisms and technological infrastructures. At its core, network science shifts the focus from individual entities to the relationships between them, asserting that the structure of these relationships often determines system-level properties—such as robustness, information flow, and collective behavior—more profoundly than the attributes of the nodes themselves. This chapter establishes the epistemological groundwork of the field, tracing its evolution from classical graph theory and early sociological experiments to modern complex systems theory. We will define the fundamental terminology—nodes, edges, ties, and graphs—distinguish between key network typologies (undirected versus directed, weighted versus unweighted, static versus dynamic), and introduce essential metrics (degree, density, path length, clustering coefficient) that quantify network structure. Historical experiments, most notably Milgram’s “small-world” study and Moreno’s sociograms, will be critically analyzed not only for their findings but also for their methodological innovations and limitations. The chapter concludes by articulating the “network perspective” as a powerful interdisciplinary lens, illustrating its application through contemporary examples from neuroscience (the human connectome) and infrastructure systems (power grids). By the end, readers will possess a concrete understanding of the basic concepts and historical context necessary to engage with the more specialized network phenomena explored in subsequent chapters.

Historical Foundations: From Graph Theory to Social Networks

The intellectual lineage of network science is remarkably multidisciplinary, with roots extending into mathematics, sociology, anthropology, and physics. Understanding this history is crucial because it reveals how the same abstract concept—a set of connections—was independently rediscovered and refined in different domains, each contributing unique theoretical tools and empirical insights. The formal study of networks began in the realm of pure mathematics with graph theory, a field concerned with the properties of abstract structures consisting of vertices (nodes) connected by edges (links). However, the systematic application of these ideas to real-world systems emerged primarily through sociology in the early twentieth century. The convergence of these two traditions in the late twentieth century, catalyzed by advances in computing and the rise of complex systems science, gave birth to modern network science as a distinct discipline.

Graph Theory: The Mathematical Abstraction

The story conventionally starts with the Seven Bridges of Königsberg problem, posed in the eighteenth century and solved by Leonhard Euler in 1736. The city of Königsberg (now Kaliningrad) was set on the Pregel River and included two large islands connected to each other and the mainland by seven bridges. The puzzle asked whether it was possible to devise a walk through the city that would cross each bridge exactly once and return to the starting point. Euler proved that no such walk existed by abstracting the problem: he represented each landmass as a point (node) and each bridge as a line (edge) connecting two points. His critical insight was that the solution depended solely on the number of edges meeting at each node (the degree of the node), not on the geometric layout of the bridges. This abstraction is considered the first theorem of graph theory and the first formal proof in the theory of networks. Euler’s work established that networks could be studied through their connectivity patterns alone, independent of spatial embedding—a principle that underpins all modern network analysis.

Throughout the nineteenth and early twentieth centuries, graph theory developed as a branch of mathematics, with significant contributions from mathematicians such as Augustin-Louis Cauchy, William Rowan Hamilton, and Arthur Cayley. Hamilton popularized the “Icosian Game” in 1857, which involved finding a Hamiltonian cycle—a path visiting each node exactly once—on a dodecahedron. Cayley, a founder of modern graph theory, studied trees (connected acyclic graphs) and enumerated different types of molecular structures, linking graph theory to chemistry. The field remained largely theoretical, concerned with existence proofs, enumeration, and combinatorial properties, with little attention to empirical systems.

A major turning point came with the work of Paul Erdős and Alfréd Rényi in the late 1950s and early 1960s. They introduced the concept of random graphs, specifically the $G(n, p)$ model, where a graph is constructed by starting with a set of $n$ isolated nodes and then adding each possible edge with probability $p$. This seemingly simple model yielded profound insights. Erdős and Rényi proved that as $n$ grows large, random graphs undergo a phase transition: there exists a critical threshold $p_c = 1/n$ such that for $p < p_c$, the graph consists of many small isolated components, while for $p > p_c$, a giant component emerges that includes a positive fraction of all nodes. This work provided the first rigorous mathematical framework for understanding how global connectivity arises from local random connections and laid the foundation for the probabilistic study of networks. However, the Erdős–Rényi model assumed all nodes are equivalent and all edges are equally likely—assumptions that would later be challenged by empirical studies of real-world networks.

Sociology: From Social Atoms to Relational Structures

While mathematicians were abstracting graphs, sociologists were developing parallel ideas to understand human societies. The pivotal figure here is Jacob Moreno, a Romanian-American psychologist and psychiatrist. In the early 1930s, Moreno introduced sociometry—the quantitative study of social relationships—and invented the sociogram, a visual representation of social structures using points and lines. His 1934 book *Who Shall Survive?* presented the first systematic use of network diagrams to map interpersonal choices in a group of schoolchildren. Moreno’s innovation was methodological and epistemological: he argued that social phenomena should be studied at the level of dyadic relationships (ties between individuals) rather than at the level of isolated individuals (“social atoms”). This relational approach fundamentally challenged the dominant individualistic paradigm in social psychology.

Moreno’s sociograms, though primitive by today’s standards, revealed patterns such as reciprocity (mutual choices), transitivity (if A chooses B and B chooses C, then A may choose C), and the existence of isolates and stars. He also developed early metrics, such as choice reciprocity and network density. However, sociometry remained a niche methodology for decades, partly due to the labor-intensive nature of data collection and the lack of computational tools for analyzing large networks.

The next major step came from the Harvard School of social relations in the 1940s and 1950s, particularly through the work of George Homans and Harrison White. Homans’s *The Human Group* (1950) emphasized that group behavior could be explained by patterns of interaction, not just individual attitudes. White, along with his students, advanced the use of mathematical models in sociology. In the 1960s and 1970s, social network analysis (SNA) emerged as a coherent paradigm, largely through the efforts of researchers like Stanley Milgram, Mark Granovetter, and Linton Freeman at institutions such as Harvard, MIT, and the University of California, Irvine. They developed key concepts—structural equivalence, role analysis, blockmodeling—and