Left identity and complementary negative directions of the covering scale commutator are not leftover interpolating sequential type

Let E_n be a named covering-class embedding and T_n a named covering-law stepper, both column-stochastic, and set Omega_n = E_n T_n - Tₙ₋₁ Eₙ₋₁ with S = sym(Omega_n). Then 1^T Omega_n = 0 for every n >= 2, hence 1^T S 1 = 0. The identity is not Omega_n 1 = 0 and is not S 1 = 0. On the pair (maximal deletion, port-resolved) the Sylvester inertia of S is (n_-, n_0, n_+) = (0,1,0), (1,0,1), (1,0,3), (1,0,10), (1,0,32), (2,0,116), (4,0,467), (8,0,2118) at n = 2 through 9. Negative eigenvalues appear at size 3, form a four-size plateau of index 1 and nullity 0 on {3, 4, 5, 6}, and then follow the look n_- = 2ⁿ⁻⁶ through size 9. On that plateau the form Q_n(x,y) = x^T S y is a Lorentzian quadratic form on the covering-class coordinate space, and the all-ones vector is null. It is not a spacetime metric. Haynsworth additivity holds on that pair from size 3. Binary doubling remains a look, as in Companion 49 Theorem 5. The upward stepper T sends the previous negative space injectively into the next coordinate space on the tested sizes. Rank forbids that map from doubling the negative dimension by itself. Principal cosines of that transport are 0.983, 0.943, 0.916 at sizes 4, 5, 6 and 0.948 at size 7; at size 8 the overlap sum of squares is 1.80 of rank 2 with Frobenius residual 0.319 and squared energy retention about 0.90. That geography is injective subspace persistence, not a parallel-transport connection. The complementary new space inside the next negative eigenspace has dimension 1 at size 7, dimension 2 at size 8, and dimension 4 at size 9. At size 9 the four principal cosines of that transport are 0.957, 0.951, 0.945, 0.941 with overlap 3.598 of rank 4 and energy retention 0.900. For each class-index i the coordinate vector u^(i) = S_ii 1 - 2 (S 1)_i e_i is null whenever 1 is null; at size 3 the two formulas collapse to one extra null vector independent of 1. That family is a rewrite of the left identity, not a stereographic projection of a sphere. Named covering-class Hasse features and birth-step schedule features do not span the size-7 complement. The unique paired-join class of Companion 49 Theorem 4 carries energy fraction 0.096 of that complement, not a claimed 22.66 percent of a secondary mode of S. Write F_n^(mu) = 1/2 (Omega_n M_n + M_n Omega_n^T). Then 1^T F 1 = 0 for every n >= 2, as a corollary of the left identity of Omega. On (maximal deletion, port-resolved) the flux inertia is n_-(F) = 0,1,1,1,1,1,3 at sizes 3 through 9: a five-size window n_-=1 on {4,5,6,7,8} with n_0=0, so Q_n^(F) is a Lorentzian quadratic form on RW_n at those sizes, not a spacetime metric, then three negatives at size 9. On (covering-order reverse, uniform-legal) the same five-size window holds, then two negatives at size 9. Those are statements about F, not about S. On that same window the unique negative direction of F has principal cosines 0.912, 0.891, 0.884, 0.893 under T at the successive steps of (maximal deletion, port-resolved); that geography is a look, not a connection. On (covering-order reverse, path-entropy) at size 7 the second negative mode of F carries energy fraction 0.227 on the paired-join class, which has eight linear extensions, not one. For a finite poset P with unique minimum 0, the Hasse-component count c_0(P) of P minus {0} is a rooted-poset invariant of covering-process orders. If every non-minimum vertex has a unique Hasse parent, then c_0(P) equals the number of covers of 0, and elements from distinct components have down-set intersection {0}. On a covering-process order, Companion 32 Theorem 1 forces c_0 in {1,2}. The classes with c_0=2 are unordered gluings of two covering-process orders along a fresh minimum, and the deposited cardinalities |W_n| give the counts 0,1,1,3,6,18,52,183,699 at sizes 2 through 10. It is not a spacetime split. Among the six named full-support pairs of the flux census, the Lorentzian plateau of S on {3,4,5,6} is unique to (maximal deletion, port-resolved), and the Lorentzian window of F on {4,...,8} is unique to the two pairs that carry that window. Occupying leftover interpolating sequential type by the left identity, by the Lorentzian bilinear form, by the complementary split, by the coordinate-null family, by a five-size window of F, by c_0, or by uniqueness of those types on the six named pairs is too loose. Remaining-never because those objects fail leftover interpolant independence is too tight. A spacetime Lorentz metric, a spacetime stereographic projection, a parallel-transport connection, a renormalisation-group flow, an identification of the complementary space with ker(E), and an overlap-law Phase-II theorem at probability one-third are not theorems of this note. Covering histories are finite posets of the mathematical type of finite causal sets. They are not identified with physical spacetime. This note does not replace the parent programme manuscript, and it is not a new version of Companions 13-49.