\section{The coherent factor and its transpose pencil} \label{sec:coherent-transpose} Shift the chosen reduced Jacobi root to zero. In this section its nonzero rank is $\nu\in\{7,8\}$, and $k=16-\nu$. Let $\Lambda$ be the diagonal matrix of the nonzero roots, with multiplicities, and put $H=\Lambda^{-1}$. Define the polarization by \[ \mathcal B(X,Y)=\tfrac12(\Jac_{X+Y}-\Jac_X-\Jac_Y). \] For a real unit anchor $X$, choose orthonormal frames $M_X:\R^\nu\to E_X=\operatorname{Im}\Jac_X$ and $K_X:\R^k\to\ker\Jac_X$, with $\Jac_X=M_X\Lambda M_X^T$. \subsection{Construction and normalization} The generic-pair and two-plane factorization statements in \cite[Section~4, Lemmas~2(1) and~3]{Nik2005} require $n\ge2\nu$, which holds here, including $\nu=8$. For a generic orthonormal pair $(X,Y)$ with disjoint Jacobi images, fixing $M_X$ uniquely determines a partner $M_Y$ for which \begin{equation}\label{eq:cp-two-plane} \Jac_{sX+tY}=(sM_X+tM_Y)\Lambda(sM_X+tM_Y)^T. \end{equation} We use only this two-plane result, not the three-plane construction in that source, whose stronger dimension hypothesis does not hold here. Choose $X$ in the nonempty open set of anchors for which the generic pair construction is available, and extend in $Y\in X^\perp$ by homogeneity. Write $M_Y=M_XF(Y)+K_XP(Y)$. Comparing the mixed coefficient in \eqref{eq:cp-two-plane} gives the actual lower block \begin{equation}\label{eq:cp-lower} P(Y)=K_X^T\,2\mathcal B(X,Y)M_X\Lambda^{-1}. \end{equation} In particular $P$ is linear; the right factor $\Lambda^{-1}$ is part of its normalization. The other block identities are \begin{align} F\Lambda+\Lambda F^T&=A(Y):=M_X^T2\mathcal B(X,Y)M_X, \label{eq:cp-mixed}\\ P\Lambda F^T&=K_X^T\Jac_Y M_X,\label{eq:cp-cross}\\ F\Lambda F^T&=M_X^T\Jac_Y M_X,\qquad P\Lambda P^T=K_X^T\Jac_Y K_X.\label{eq:cp-diagonal} \end{align} The generic disjointness of the images makes $P$ injective there. Solving \eqref{eq:cp-cross} with a rational left inverse of $P$ therefore makes $F$ rational and homogeneous of degree one. All the displayed equations extend as rational identities. Consequently \begin{equation}\label{eq:cp-global-factor} M_{sX+Y}=sM_X+M_XF(Y)+K_XP(Y),\qquad \Jac_{sX+Y}=M_{sX+Y}\Lambda M_{sX+Y}^T. \end{equation} \begin{lemma}\label{lem:cp-pole-free} If $F$ has no pole along any complex hypersurface in $X^\perp$, then \eqref{eq:cp-global-factor} is a global linear factor. In that event the tensor is Clifford. \end{lemma} \begin{proof} In the polynomial unique factorization domain on $X^\perp$, each entry of a rational function with no divisorial pole has constant reduced denominator. Thus $F$ is polynomial. Its degree-one homogeneity makes it linear. Equation~\eqref{eq:cp-global-factor} is then a polynomial identity on all sixteen variables. At every nonzero real vector its factor is injective because the Jacobi operator has rank $\nu$. Proposition~\ref{prop:gramconditional}(1) applies. \end{proof} \subsection{Kernel directions give genuine Jacobi factors} Extend $P$ linearly to $V=\R^{16}$ by $P(X)=0$. For $g\in\R^k$ define $N_g:\R^\nu\to V$ by \begin{equation}\label{eq:cp-N} N_g^T Y=P(Y)^Tg\qquad(Y\in V). \end{equation} This is linear in $g$. For a coefficient $a\in\R^\nu$, define the transpose pencil $B(a):\R^k\to V$ by $B(a)g=N_ga$. \begin{lemma}\label{lem:cp-L20} For every real $g$, \begin{equation}\label{eq:cp-kernel-factor} \Jac_{K_Xg}=N_g\Lambda N_g^T. \end{equation} If $g\ne0$, the map $N_g$ is injective. If $a\ne0$ is real, the map $B(a)$ is injective, equivalently $Y\mapsto P(Y)a$ is onto $\R^k$. There are no nonzero fixed real coefficient space $U$ and proper fixed real subspace $T\subset\R^k$ such that $P(Y)U\subset T$ for all $Y$. \end{lemma} \begin{proof} For $z=K_Xg$, the scalar identity \[ \langle\Jac_Yz,z\rangle =(P(Y)^Tg)^T\Lambda(P(Y)^Tg) =\langle\Jac_zY,Y\rangle \] follows from the coherent factor and sectional symmetry. Polarizing in $Y$ proves \eqref{eq:cp-kernel-factor}. Its left side has rank $\nu$ for nonzero real $g$, so $N_g$ has full column rank. If $B(a)g=N_ga=0$ for nonzero real $a$, this injectivity forces $g=0$. Transposition proves the surjectivity claim. Finally, choose nonzero $g\perp T$. The proposed containment would make $N_g$ annihilate $U$, contradicting its injectivity. \end{proof} Let $g_X$ be the unit vector with $K_Xg_X=X$. The radial and skew identities used below are \begin{equation}\label{eq:cp-radial-skew} B(a)g_X=-M_Xa,\qquad K_X^TB(a)\text{ is skew-symmetric}, \qquad\operatorname{Im}B(a)\subset X^\perp. \end{equation} Indeed injectivity of the coherent factor and $\Jac_zz=0$ give $M_z^Tz=0$ on its domain. The coefficient of $s$ after writing $z=sX+Y$ gives $P(Y)^Tg_X=-M_X^TY$, proving the first identity. The last follows from $P(X)=0$. For the middle identity, \eqref{eq:cp-kernel-factor} and $\Jac_{K_Xg}K_Xg=0$ imply $N_g^TK_Xg=0$ at real $g\ne0$. Hence $\langle g,K_X^TB(a)g\rangle=0$ for all real $g,a$; polarization in $g$ proves skewness. \begin{lemma}\label{lem:cp-complex-ranks} For every nonzero complex coefficient $a$, the kernel of $B(a)$ is totally isotropic in $\mathbb C^k$ for the standard bilinear form. Writing $s=\rank B(a)$, one has \[ s\ge\lceil k/2\rceil,\qquad \rank(B(a)B(a)^T)=2s-k. \] \end{lemma} \begin{proof} If $0\ne g\in\ker B(a)$ and $g^Tg\ne0$, the continued Osserman characteristic and minimal-polynomial identities make $\Jac_{K_Xg}$ diagonalizable of rank $\nu$. Equation~\eqref{eq:cp-kernel-factor}, continued polynomially in $g$, then forces $N_g$ to have rank $\nu$, contrary to $N_ga=0$. Thus every vector of $\ker B(a)$ is null. Polarization makes that kernel totally isotropic, so its dimension is at most $\lfloor k/2\rfloor$. If $L=\ker B(a)$, then $\operatorname{Im}B(a)^T=L^\perp$ and $L\subset L^\perp$. The image of $B(a)B(a)^T$ therefore has dimension $\dim L^\perp-\dim L=2s-k$. \end{proof} \subsection{The local pole budget} Put $F_0=\tfrac12 A\Lambda^{-1}$, with $A$ from \eqref{eq:cp-mixed}, and set $\Omega=(F-F_0)\Lambda$. Then $\Omega$ is skew-symmetric and homogeneous of degree one. Equations~\eqref{eq:cp-cross}--\eqref{eq:cp-diagonal} imply that $P\Omega$ is polynomial and that the negative-order terms of $\Omega H\Omega^T$ are cancelled only by terms of at most the pole order of $\Omega$. \begin{lemma}\label{lem:cp-dvr-budget} Let $f$ be an irreducible polynomial defining a genuine pole of $\Omega$, of order $d>0$. Every maximal $\nu$-minor of $P$ is divisible by $f^{2d}$. If $d=1$ and the residue has rank $r$, every such minor is divisible by $f^r$. Consequently, if $D$ is the reduced common denominator of $\Omega$, then $D^2$ divides every maximal minor of $P$, and $2\deg D\le\nu$. \end{lemma} \begin{proof} Work over the discrete valuation ring obtained by localizing the polynomial ring at $(f)$, with uniformizer $\pi=f$. An alternating matrix over its fraction field is put, by a unimodular congruence, into two-by-two blocks $a_j\left[\begin{smallmatrix}0&1\\-1&0\end{smallmatrix}\right]$ and a zero block. One obtains this form by choosing an entry of least valuation as pivot and clearing its two rows and columns; all elimination coefficients are integral. Repeating gives the required congruence. Write the negative block valuations as $-d_j$, with $d_j>0$. Because $P\Omega$ is integral, the corresponding two columns of the unimodularly transformed $P$ are divisible by $\pi^{d_j}$. Thus every maximal minor has valuation at least $2\sum_jd_j$. The largest $d_j$ is the pole order $d$, proving the first bound. At a simple pole the residue rank is twice the number of negative blocks, giving the second bound. Unimodular column changes preserve the maximal-minor ideal. Distinct irreducible factors multiply in the polynomial ring, proving the assertion for $D$. At least one maximal minor is nonzero, since $P$ is generically injective; that minor has degree $\nu$. \end{proof} If $W$ is the highest Laurent coefficient of $\Omega$ at a pole, the leading terms of the same identities give \begin{equation}\label{eq:cp-residue} PW=0,\qquad W^T=-W,\qquad WHW^T=0 \end{equation} over the residue field. In particular its nonzero rank is even, and its image is $H$-isotropic. The upper-block identity gives the last equation because its order $2d$ term has no other possible cancellation. In the square case $\nu=8$, the budget reads $D^2\mid\det P$ and $\deg D\le4$. The real factor types and the subsequent support reductions are separate steps; they do not follow from this degree bound alone.