Within the depth-psychology corpus broadly construed, ‘network’ operates across at least three analytically distinct registers that nonetheless converge on a shared insight: that mind, affect, and selfhood are constitutively relational rather than localised. In cognitive neuroscience and affective science—represented here by Barrett, Menon, Craig, Carhart-Harris, and Paulus—the term designates large-scale functional brain systems (default mode network, salience network, central executive network, interoceptive network) whose reciprocal coupling and dynamic switching underlie emotion construction, predictive processing, and self-regulation. Barrett’s account of the interoceptive network as the substrate for emotion construction is paradigmatic of this position. Menon’s tripartite network model formalises the switching logic that has become foundational for understanding psychopathology. A second register, represented by Thompson, treats the network as a figure for autopoietic biological organisation: the metabolic network of the cell that recursively regenerates both its components and its boundary is the minimal prototype for mind-in-life. A third, more classical connectionist usage appears in Thompson’s account of artificial neural networks as candidate models of cognition. Allan’s cognitive-linguistic usage—semantic category networks—bridges biological and computational concerns through the notion of polysemy. The central tension is between network-as-substrate (fixed functional architecture) and network-as-process (dynamically reconfigured, entropy-sensitive, state-dependent). Carhart-Harris’s entropic-brain thesis sharpens this tension by treating increased network entropy as constitutive of primary states of consciousness. Across all registers, network thinking displaces the classical location of mental functions in discrete regions, insisting instead on distributed, emergent, and relational organisation.