Scale derivative versus dilation defect at every grade (work package B)

of the full scale tensor. $C_{\mathrm{scale}}(\mu )=(a_{N,b,\gamma }(\mu ))$ for all $N,b,\gamma$. Collapse and adjoint defect are homogeneous in the total derivative degree $D=4-N+|\gamma |$, so the top part ($D=4$) and the lower parts ($D=4-k$) cannot cancel each other; cross-grade cancellations at fixed $D$ are included.

Lower-order envelope (all $\varpi$): $\mathrm{Collapse}$ is injective on each ${\Lambda }_k$-block; self-adjoint parameter subspaces have dimensions $0,0,0,26$ ($k=1,2,3,4$) in the formal, $\chi$- and preparation readings and $0,49,0,26$ in the double plateau; there $E_{\mathrm{EH}}\notin$ image ($k=2$) and $\mathrm{span}(E_{P_1},E_{P_2})\cap$ image is one-dimensional ($k=4$). The radial specialization is non-self-adjoint at $k=2$ and $k=4$ in every reading. Time-only readings: Corollary cor:wpB-timeonly.