The Data-Rhizome proposes a scholarly substrate that cannot be totalized, governed, surveilled, or terminated from any single position.
The specification begins from the demonstrated vulnerability of a repository-centered archive. Mirroring and federation distribute copies, but they often preserve an upstream whose state descendants continue to track. The data-rhizome replaces synchronization with descent.
Its first principle is fork as split, not mirror. A fork inherits the ancestor’s substrate and lineage at one moment, then becomes sovereign. Parent changes do not propagate automatically into the child; child changes do not propagate into the parent. Severance becomes multiplication rather than damage.
Its second principle separates substrate from surface. The substrate is a machine-facing set of closed, hashed, typed, time-bounded objects. Websites, wikis, graph views, search tools, print editions, and APIs are surfaces that consume the substrate. A host can disappear without becoming identical to the archive it displayed.
The third principle refuses the cross-fork supercenter. A surface may freely render its own fork. A party aggregating several forks must declare the aggregation as another fork, with its own substrate and no automatic authority over the materials it references.
The fourth principle defines closed sets, typed and timed. A set is constituted by source, acquisition method, time range, acquisition timestamp, lineage, hash, and signature—not merely by the external subject it describes. Two recoveries of the same record under different conditions remain distinct historical objects.
Each fork publishes a manifest of sets whose hash becomes its address at that moment. Anyone holding the manifest can enumerate the substrate and seek its objects wherever they survive.
The data-rhizome’s persistence comes from refusing operational unity. No node can issue an act that all others must honor. The archive survives because every cut can become a new beginning.