[DSRP Evidence](https://dsrpevidence.org/) # Is Network Theory Made of DSRP? ## Details **Authors** Cabrera, Derek **Year** 2026 **Publisher** Journal of Systems Thinking **Kind of work** article **Discipline** Network Science [Read it at the publisher](https://doi.org/10.54120/jost.v26i1.dc6ntsd6) 10.54120/jost.v26i1.dc6ntsd6 ## In authors' words ### Abstract Network science is among the most portable representational traditions in modern science, and its primitives are among the fewest. A network is a set of nodes and a set of edges; from that pair the field derives degree, paths, components, clustering, centrality, communities, robustness, and the growth models that generate them. The abstraction descends from a single act performed by Euler in 1735, who solved the Königsberg bridge problem by deleting every measurable property of the city and keeping only which things exist and which are related. DSRP Theory makes a different and more elemental claim: that organization itself can be represented by four recursively entangled Patterns – Distinctions, Systems, Relationships, and Perspectives – composed of eight elemental roles and governed by four structural dynamics. This paper asks whether the constructs of network science introduce any organizational structure not already expressible by DSRP-484. Extending an earlier treatment that decomposed the node and the edge and derived counting heuristics for each Pattern, we execute a pre-registered structural-remainder test against a candidate set of thirty-five constructs drawn exclusively from the field's own vocabulary – Newman's canonical survey, Barabási's textbook, and Euler's founding paper – under the same fixed extraction rule and the same information/organization decision rule applied to system dynamics in the companion run. All thirty-five decompose, with no residual: nodes as Distinctions whose identity is a ledger of what they are and are not; edges as Relationships with action-reaction substructure, themselves reifiable as objects; paths as composed directed Relationships; components as Systems reconstructed from Relationships under a rule-Perspective; centrality as a Perspective on importance; preferential attachment as a reinforcing feedback loop closed through the attaching node's Perspective. The analysis also yields a finding the test was not designed to produce. Six of network science's load-bearing constructs – betweenness centrality, the boundary specification problem, the small-world rewiring effect, the scale-free disproportion, the isolated but contained member, and the hyperedge – work around part-whole and perspectival structure that the formalism's own primitives cannot represent. Every workaround succeeds; none revises the primitives. More broadly, the field's extensions collectively reconstruct the seven elemental roles Euler's traversal abstraction left implicit, often in combination, without introducing an organizational primitive outside DSRP-484. The field's signature findings run on organization its grammar denies. One legitimate network construct requiring a fifth irreducible Pattern, a ninth elemental role, or another irreducible structural dynamic would falsify the claim. None was found. Taken with the system dynamics run, ninety-five constructs across two independent traditions decompose to the same four Patterns – never a fifth, never a different four. ### What they set out to do (purpose) Second run in the structural-containment series. Asks whether network science introduces any organizational structure outside DSRP-484, using the protocol fixed in the system dynamics companion with only the nominating sources changed. ### What they found (results) Thirty-five canonical constructs entered the candidate set; ten entries were excluded and logged by category. All thirty-five decompose with no residual – four had been decomposed in the precursor chapter and one partly, twenty-five required new decomposition, two are composites of already-decomposed elements, and three are purely informational. The residual category X is empty. A second result the test was not designed to produce: Euler kept two of the eight elemental roles (identity, and a generic Relationship with no roles inside it) and left seven implicit, and the field's later extensions – boundary rules, components, communities, hypergraphs, bipartite transformations, directed and typed edges, multiplex and hierarchical networks, line graphs, centrality measures, higher-order models – repeatedly reconstruct combinations of those same seven without adding a new primitive. Cabrera reads that repetition as the evidence: apparently unrelated higher-order inventions keep solving versions of the same lower-order omissions. He states nineteen prospective predictions and fifteen falsifiers, and reports the claim as exposed and unfalsified on this run rather than confirmed. ## Commentary ### In short Our reading: the historical argument is what this paper adds over the SD run. The decomposition itself is performed by an analyst who knows DSRP, and that is a fair objection; the extensions were not – they were built over two centuries by network scientists solving field-specific problems with no interest in DSRP. That the extensions converge on the omitted roles is evidence that does not depend on the analyst's prior commitment. The hypergraph is the cleanest case: the phenomenon says part-whole, the grammar has no primitive for it, so the field calls the whole an edge and has been building the missing role for two decades under a name describing it as a kind of edge. ### In more detail Pre-registered nominators: Newman's "The Structure and Function of Complex Networks" (every term in its terminology box, every property and model given a section-level definition), Barabási's Network Science for terms Newman does not define, and Euler's 1741 paper for the constructs of the field's origin act. Wasserman and Faust serve as the robustness check. The earlier DSRP treatment of networks was deliberately excluded as a nominator: the DSRP side of a containment test must not choose what the other side is asked to explain. Exclusion categories match the SD run and every exclusion is logged – graph traversal and search algorithms, generating function methods, mean-field and master-equation approximations, exponential random graph estimation (mathematical technique); community-detection algorithms and network sampling (research practice); adjacency matrices and layout conventions (representation artifact); "complex system" in general usage (general term); epidemiological compartment models (imported theory). Synonyms the sources themselves equate – vertex/node/site/actor, edge/link/bond/tie, cumulative advantage/preferential attachment – are merged with aliases recorded. Section 17 gives the full thirty-five-row decomposition table with a residual column; the supplementary instrument records each ruling with its verbatim source definition, so any single row can be contested. **Patterns it shows** D, S, R, P Shows the structure is there, Shows it makes a difference **How to cite this** Cabrera, Derek (2026). Is Network Theory Made of DSRP?. Journal of Systems Thinking.