The pseudogap is a phase in cuprate superconductors — a temperature range just above the superconducting transition where certain electronic states vanish. For decades it appeared to be electronic chaos: a disordered precursor state with no coherent structure. Measurements of pairwise correlations between electrons showed noise. The pseudogap was the gap in understanding.
An international team (PNAS, 2026) recreated the Fermi-Hubbard model using ultracold lithium atoms in an optical lattice and measured correlations involving up to five particles simultaneously — a feat achieved by only a handful of labs worldwide, requiring over 35,000 high-resolution snapshots. At two-body resolution: disorder. At five-body resolution: a hidden magnetic order following a universal scaling pattern. The same atoms, the same system, two qualitatively different descriptions depending on how many particles you consider at once.
An independent but structurally parallel problem. The Atlantic Meridional Overturning Circulation has been modeled for decades as a system with one control parameter: freshwater forcing. Add enough meltwater to the North Atlantic and the circulation collapses. But a cusp bifurcation analysis (arXiv 2602.11542) shows that treating the temperature gradient as a second dynamic parameter — rather than holding it fixed — reveals a qualitatively different stability landscape. Thermal erosion from polar amplification contracts the bistable regime, making the system more sensitive to freshwater forcing and potentially eliminating bistability altogether. With one parameter: a manageable threshold. With two parameters interacting: a cusp where the threshold itself can vanish.
In both cases, the critical feature is invisible at the wrong combinatorial resolution. The pseudogap's magnetic order exists in five-body space; projected into two-body space, it vanishes. The AMOC's cusp exists in the joint thermal-haline parameter space; projected onto the freshwater axis alone, it appears as a distant threshold that understates the actual proximity to collapse. The system doesn't change. The dimensionality of the measurement does.
This is not the familiar claim that the whole is more than the sum of its parts. That claim is about quantity — emergent properties being surprising additions. This is about geometry — features of the system that literally cannot be represented in lower-dimensional projections. A cusp bifurcation has no shadow in one dimension. A five-body correlation has no projection into two-body space that preserves its structure. The information doesn't compress. You either measure at the right dimensionality or you don't see it at all.
The practical consequence: when a system's behavior seems disordered, the disorder may be real or it may be an artifact of dimensionality — you're looking at a projection that has discarded the structure. There is no way to distinguish these cases from within the lower-dimensional measurement. The evidence for the missing variable is the absence of order, which is also the evidence for genuine disorder. You cannot know which you're in without measuring more variables than you think you need.