Scalar lensing makes only E-modes, so shear B-modes are used as pure systematics alarms: CFHTLenS detected them without tracing their origin, KiDS required scale cuts, and DES and HSC pass at thresholds that absorb known residuals. Every E-mode cosmology, including the S₈ measurements, inherits the systematic floor the B-tests merely bound, and any genuine B-mode physics is definitionally filed under instrument error.
The null-test logic encodes the model: density perturbations are scalar, gravity is standard, so curl patterns cannot be cosmological. The assumption forecloses the measurement: a real B-mode source would be cut, calibrated, and marginalized away as PSF residue by pipelines built never to believe it.
SCT's lensing sky is not purely scalar: the coherent mesh contribution to the potential (P50, P51, P53) carries vector character, since the superposition of comoving sources encodes their shared bulk motion, gravitomagnetic terms that scalar-only bookkeeping omits, and the inherited rotational structure of the deposition (P31, P32) organizes those contributions coherently across large scales. Coherent vector components in the deflection field generate genuine curl: a small physical B-mode floor, survey-independent, correlated with large-scale structure orientation rather than with any instrument's optics. In epistemic honesty, the amplitude derivation is flagged as outstanding theoretical work; the structural prediction is that the floor exists, is shared, and is sky-correlated.
The discriminating signature divides cleanly: systematics are instrument-specific, uncorrelated between Euclid, Rubin, KiDS, and DES at fixed sky, while a mesh-sourced floor must agree between instruments observing the same fields and correlate with the web's J-alignment geography (the spin-filament and polarization-alignment family). The persistent cross-survey B-mode residue the null tests keep bounding is, on this reading, the first faint look at the mesh's vector sector, and the same omission that surfaces as the A_lens smoothing excess in the CMB.
This is the same coherent-superposition physics behind the lensing-amplitude family, read in the curl channel. There is no need to define the discovery away before measuring its cross-survey coherence.
The test is correlation: Euclid and Rubin measuring B-mode residuals on shared sky and finding them instrument-specific, uncorrelated between surveys and with structure orientation, would establish the systematics reading and remove the vector-mesh floor at current sensitivity. A confirmed cross-survey, structure-correlated B-mode floor with the J-alignment geography, conversely, has no standard-model source; its absence at Euclid-Rubin precision bounds the mesh's vector sector toward zero and weakens this entry without touching the scalar coherence results.