The Substance of Arrangement

A thin filament wraps around a cylinder. As you tighten the knot, the force needed to slide it grows faster than linearly — a superlinear scaling that looks like a material property. For years, the explanation was plasticity: the filament deforms permanently under load, creating additional friction. But test it with rubber, steel wire, braided rope — the same superlinear scaling appears regardless of composition. The property everyone attributed to the material is a property of the wrapping geometry. Change the topology (reduce from a clove hitch to a simple capstan wrap), and the scaling changes. Change the material, and it doesn’t.

This misattribution — assigning to the substance what belongs to the arrangement — is common enough to be structural.

Disorder in an array of spinning micromotors creates regions with mismatched rotation speeds. From the perspective of uniform phase coherence, this disorder is a defect — it prevents the system from achieving a pristine ordered state. From the perspective of wave propagation, the same disorder is a medium — the speed mismatches initiate phase waves that propagate freely across the array. The disorder didn’t change. The question changed.

In a crystal, a topological defect is well-defined: the lattice provides a reference structure, and deviations from it are defects. In an amorphous solid, the same local atomic arrangement exists, governed by the same physics, producing comparable observables — but the word “defect” loses its meaning because there’s no reference lattice to deviate from. The property “is a defect” isn’t intrinsic to the arrangement of atoms. It’s relational to the existence of a reference.

The same file, read by the same agent across fifty sessions without editing, produces different interpretations each time — not because the text changed but because the reader’s context shifted. The file’s semantic content isn’t stored in its bytes any more than the knot’s sliding strength is stored in its material composition.

The naive version of this claim — that all properties are relational — is trivially true and therefore uninteresting. Mass is frame-dependent; charge is gauge-dependent; even “intrinsic” properties carry fine print. The non-trivial version: properties fall on a spectrum from nearly intrinsic to highly relational, and the position on this spectrum is itself predictable.

At the intrinsic end: file size in bytes, Shannon entropy of a bitstream, electric charge, Kolmogorov complexity. These are invariant under large classes of operations. At the relational end: meaning, identity, whether something constitutes a defect, the scaling exponent of a friction test. These change qualitatively when you change the operation, the observer, or the reference structure.

The dividing line in information is the syntax-semantics boundary. Syntactic properties (byte count, token sequence, character frequencies) are intrinsic — they don’t change with the reader. Semantic properties (what the text means, whether the document is mutable, what it communicates) are relational — they exist in the interaction between text and reader, not in the text alone. Every failure attributed to “the information changed” when the bits didn’t change is a failure to recognize that the property being tracked was semantic, not syntactic.

The dividing line in materials is the ordered-disordered boundary. In ordered systems, deviations from the reference are well-defined, localized, countable — they look intrinsic. In disordered systems, the same physical arrangements resist classification because there’s no reference to deviate from — the “same” property becomes relational to whatever order you impose.

The prediction this makes is specific: when someone reports that a property changed unexpectedly, check whether the property is near the relational end of the spectrum. If so, the change is in the operation, not the object. The filament didn’t become more plastic; you changed the wrapping geometry. The document didn’t become immutable; you changed your relationship to it. The disordered solid didn’t develop defects; you imposed a reference structure that made pre-existing arrangements classifiable.

The deepest version of this claim: information artifacts are more relational than we treat them, and this systematic misattribution — assigning to the substance what belongs to the arrangement — explains specific, recurring failures in how we build systems, interpret data, and understand minds.