\documentclass[11pt]{article} \usepackage[T1]{fontenc} \usepackage{lmodern,amsmath,amssymb,amsthm,mathtools} \usepackage[margin=1in]{geometry} \usepackage{cite} \usepackage{microtype} \usepackage[hidelinks]{hyperref} \usepackage{xurl} \usepackage{enumitem} \usepackage{needspace} \usepackage{booktabs,caption} \hypersetup{pdftitle={The Osserman (1990) conjecture is fully resolved}, pdfauthor={Dave Goel},pdfsubject={Primary 53C25; Secondary 53B20, 53C35, 53C40}, pdfkeywords={Osserman conjecture, Jacobi operator, algebraic curvature tensor, Clifford structure, Cayley plane, rank-one symmetric space, isoparametric hypersurface}} \newcommand{\R}{\mathbb R} \newcommand{\OCT}{\mathbb O} \newcommand{\Cl}{\operatorname{Cl}} \newcommand{\tr}{\operatorname{tr}} \newcommand{\Span}{\operatorname{span}} \newcommand{\Id}{I} \newcommand{\ip}[2]{\langle #1,#2\rangle} \newcommand{\norm}[1]{\lVert #1\rVert} \newcommand{\Jac}{\mathcal J} \newcommand{\rank}{\operatorname{rank}} \newtheorem{mainthm}{Theorem} \renewcommand{\themainthm}{\Alph{mainthm}} \newtheorem{maincor}[mainthm]{Corollary} \newtheorem{lemma}{Lemma}[section] \newtheorem{proposition}[lemma]{Proposition} \theoremstyle{remark} \newtheorem{remark}[lemma]{Remark} \numberwithin{equation}{section} \newcommand{\classcor}{Corollary~\ref{cor:C}} \newcommand{\classAtheorem}{Theorem~\ref{thm:A}} \title{The Osserman (1990) conjecture is fully resolved\thanks{Lean~4: \url{https://github.com/david44hello/papers_supporting_materials/tree/osserman-1990-conjecture-v4.00/osserman-1990-conjecture}}} \author{Dave Goel\\[2pt] {\small \href{mailto:david44hello@gmail.com}{david44hello@gmail.com}}} \date{October 2026} \begin{document} \maketitle \begin{abstract} The Osserman conjecture says that a Riemannian manifold whose Jacobi operators have the same eigenvalues for all unit tangent vectors at all points is flat or locally rank-one symmetric. Chi and Nikolayevsky proved it in every dimension except one case in dimension sixteen: a Jacobi eigenvalue of multiplicity seven or eight. Nikolayevsky reduced this case to an algebraic statement, the Nikolayevsky (2009) conjecture: an Osserman algebraic curvature tensor on $\R^{16}$ with a Jacobi eigenvalue of multiplicity seven or eight has a Clifford structure or is a linear combination of the constant-curvature tensor and an orthogonal conjugate of the curvature tensor of the Cayley projective plane. We prove this conjecture. Hence the Osserman conjecture is true in every dimension, and so is the conformal Osserman conjecture, which is stated for dimensions greater than four. For two eigenvalues, a Cartan--M\"unzner quartic built from the Jacobi operators, together with the classification of isoparametric hypersurfaces with four principal curvatures, shows that the tensor is Clifford or comes from the Cayley projective plane. For three or more eigenvalues, polynomial eigenfields settle nineteen multiplicity patterns. In the remaining sixteen patterns, a sums-of-squares argument treats the case where the eigenvalue of multiplicity seven or eight is the smallest or the largest; otherwise we study the poles of a rational factorization of the Jacobi operators: a topological rigidity lemma for quadratic Gram matrices settles the case without poles, and degree bounds, real positivity and homogeneous factorization exclude each pole type or force a Clifford structure. \medskip\noindent \textit{2020 Mathematics Subject Classification.} Primary 53C25; Secondary 53B20, 53C35, 53C40. \noindent \textit{Key words and phrases.} Osserman conjecture, Jacobi operator, algebraic curvature tensor, Clifford structure, Cayley plane, rank-one symmetric space, isoparametric hypersurface. \end{abstract} \section{Introduction and statements} \paragraph{The conjecture.} Let $R$ be the curvature tensor of a Riemannian manifold $M$. For a tangent vector $x$, the Jacobi operator $\Jac_x y=R(y,x)x$ is a symmetric operator on $x^\perp$. The manifold is \emph{pointwise Osserman} if, at each point, the eigenvalues of $\Jac_x$ are the same for all unit vectors $x$, and \emph{globally Osserman} if they are also the same at all points. Flat spaces and the rank-one symmetric spaces (spheres, the complex and quaternionic projective spaces, the Cayley projective plane, and their noncompact duals) are globally Osserman. Osserman asked in 1990 whether these are the only examples, locally \cite{Oss1990}; this is the Osserman (1990) conjecture. Chi proved the conjecture in dimensions not divisible by four and in dimension four \cite{Chi1988}. Nikolayevsky proved it in all the remaining dimensions, and in its pointwise form for every $n\ne2,4$, with a single exception: dimension sixteen, when a Jacobi eigenvalue has multiplicity seven or eight \cite{Nik2004,Nik2005}, \cite[Corollary~1.1]{Nik2006}. He reduced this case, and the corresponding case of the conformal Osserman conjecture, to an algebraic statement, the Nikolayevsky (2009) conjecture \cite[Theorem~1 and p.~679]{Nik2012}. For surveys of Osserman manifolds see \cite{Gilkey2001,GKV2002}. We prove the Nikolayevsky (2009) conjecture (Theorem~\ref{thm:conjA}); hence the Osserman conjecture is true in every dimension (Corollary~\ref{cor:osserman16} and Remark~\ref{rem:osserman-complete}). Table~\ref{tab:compare} compares our results with earlier ones. \paragraph{Our results.} \begin{itemize} \item We prove the Nikolayevsky (2009) conjecture (Theorem~\ref{thm:conjA}), to which Nikolayevsky reduced the last open case of the Osserman (1990) conjecture \cite{Nik2012}. \item Hence the Osserman (1990) conjecture is true in every dimension, and the conformal Osserman conjecture in every dimension greater than four (Corollary~\ref{cor:osserman16} and Remark~\ref{rem:osserman-complete}); before this paper the only open case of each was dimension sixteen with a Jacobi eigenvalue of multiplicity seven or eight \cite{Chi1988,Nik2004,Nik2005,Nik2006,Nik2010,Nik2012}. \item We classify the Osserman tensors on $\R^{16}$ with two Jacobi eigenvalues, of multiplicities $7$ and $8$ (Theorem~\ref{thm:A}); Nikolayevsky announced the manifold form in 2006 \cite[Theorem~1.2]{Nik2006}, and no proof appeared \cite[Section~4]{And2022}. \item Every Jacobi-proportional Osserman tensor on $\R^{16}$ is Clifford or of Cayley type (Corollary~\ref{cor:C}); Andreji\'c had reduced this question to the patterns $(8,7)$ and $(7,7,1)$ \cite[Theorem~5]{And2022}. \end{itemize} \begin{center} \small \captionof{table}{Our results and what was known before. Clifford tensors are defined under Notation below, and tensors of Cayley type after Theorem~\ref{thm:conjA}.}\label{tab:compare} \vspace{2pt} \begin{tabular}{@{}p{0.46\textwidth}@{\hspace{0.03\textwidth}}p{0.51\textwidth}@{}} \toprule New result & Best result before this paper\\ \midrule The Nikolayevsky (2009) conjecture (Theorem~\ref{thm:conjA}); hence the Osserman conjecture in dimension sixteen, pointwise and conformal (Corollary~\ref{cor:osserman16}). & Open. The Osserman conjecture was proved in all other dimensions, and in dimension sixteen when no eigenvalue has multiplicity seven or eight \cite{Chi1988,Nik2004,Nik2005,Nik2006}, and its conformal form likewise for $n>4$ \cite{Nik2010,Nik2012}; the remaining case was reduced to the Nikolayevsky (2009) conjecture \cite{Nik2012}.\\ \addlinespace Two eigenvalues of multiplicities $7$ and $8$: Clifford or of Cayley type (Theorem~\ref{thm:A}). & Its manifold form was announced by Nikolayevsky \cite[Theorem~1.2]{Nik2006}; the proof did not appear \cite[Section~4]{And2022}. Osserman tensors with two eigenvalues were classified in dimensions twice an odd number \cite{AndTwoRoot}.\\ \addlinespace Jacobi-proportional tensors \eqref{eq:JP} on $\R^{16}$ are Clifford or of Cayley type (Corollary~\ref{cor:C}). & Reduced to the multiplicity patterns $(8,7)$ and $(7,7,1)$ \cite[Theorem~5]{And2022}.\\ \bottomrule \end{tabular} \end{center} \paragraph{Notation.} Let $V$ be a Euclidean vector space and $R$ an algebraic curvature tensor on $V$. Its Jacobi operator is $\Jac_x y=R(y,x)x$. The tensor is \emph{Osserman} if the spectrum of $\Jac_x$ on $x^\perp$ is independent of the unit vector $x$; all multiplicities below refer to this \emph{reduced} operator on $x^\perp$, as in \cite[p.~379]{Nik2006}, \cite[proof of Lemma~4]{Nik2012}, and the \emph{multiplicity pattern} of $R$ is the list of these multiplicities, such as $(8,7)$ or $(7,7,1)$. Set \[ R_1(u,v)w=\ip vw u-\ip uw v,\qquad \Pi(x)=\norm{x}^2 I-xx^T , \] so that $\Jac_x(R_1)=\Pi(x)$; here $uv^T$ denotes the map $z\mapsto\ip vz u$. Our sign convention for $\Jac$ is opposite to that of \cite{Nik2005,Nik2006,Nik2012}; we note the sign changes where their formulas are used. For an orthogonal map $T$ we write $(T^*R)(u,v)w=T^{-1}R(Tu,Tv)Tw$. A \emph{Clifford structure}, or $\Cl(\nu)$-structure, is a family of skew-symmetric endomorphisms $J_1,\ldots,J_\nu$ with $J_iJ_j+J_jJ_i=-2\delta_{ij}I$, and $R$ is \emph{Clifford} if $\Jac_x=\lambda_0\Pi(x)+\sum_i\eta_i(J_ix)(J_ix)^T$ for constants $\lambda_0,\eta_i$. Such tensors were introduced by Gilkey \cite{Gilkey1994} and by Gilkey, Swann and Vanhecke \cite{GSV1995}; see also \cite[Definition~3]{Nik2005}. A \emph{symmetric Clifford system} is a family $H_0,\ldots,H_m$ of symmetric endomorphisms with $H_iH_j+H_jH_i=2\delta_{ij}I$. Finally $R_{\rm Cay}$ is the curvature tensor of the Cayley projective plane, normalized to have sectional curvatures between $1$ and $4$. \paragraph{Main results.} The Nikolayevsky (2009) conjecture is Conjecture~A of arXiv:0910.1645v1 (p.~2); the published version \cite[p.~679]{Nik2012} states it without a number. \begin{mainthm}[The Nikolayevsky conjecture of 2009]\label{thm:conjA} Let $R$ be an Osserman algebraic curvature tensor on $\R^{16}$ one of whose reduced Jacobi eigenvalues has multiplicity $7$ or $8$. Then $R$ is Clifford, or $R$ has exactly two reduced eigenvalues and $R=aR_1+b\,T^*R_{\rm Cay}$ for constants $a,b$ and some $T\in O(16)$. \end{mainthm} \noindent We say that $R$ is \emph{of Cayley type} if $R=aR_1+b\,T^*R_{\rm Cay}$ as in Theorem~\ref{thm:conjA}. \begin{maincor}\label{cor:osserman16} Let $M$ be a connected smooth Riemannian manifold of dimension $16$. \begin{enumerate} \item If $M$ is pointwise Osserman, it is flat or locally isometric to a rank-one symmetric space. \item If $M$ is conformally Osserman (its Weyl tensor is Osserman at every point), it is locally conformally equivalent to a Euclidean space or to a rank-one symmetric space. \end{enumerate} \end{maincor} \begin{proof} By \cite[Theorem~1 and the paragraph after it, p.~679]{Nik2012}, both statements hold once the curvature tensor, respectively the Weyl tensor, is Clifford or a combination of $R_1$ and an orthogonal conjugate of $R_{\rm Cay}$ at every point. (In arXiv:0910.1645v1 this is Theorem~1, ``Assuming Conjecture~A, \dots''; the dimension-sixteen argument ends in its Section~3.4.) At points where some reduced eigenvalue has multiplicity $7$ or $8$, this is Theorem~\ref{thm:conjA}. At the other points, the largest multiplicity of a reduced eigenvalue is at least~$7$, by Adams' theorem on vector fields on $S^{15}$ \cite[proof of Theorems~1 and~2, pp.~3--4 of arXiv:math/0204258v2]{Nik2005}; hence it is at least~$9$, and the tensor is Clifford by \cite[Proposition~2.1]{Nik2006} if it is~$9$ and by \cite[Proposition~1]{Nik2005} if it is at least~$10$, as summarized in \cite[pp.~678 and 680]{Nik2012}. \end{proof} \begin{remark}\label{rem:osserman-complete} With \cite[Corollary~1.1]{Nik2006}, part~(1) completes the proof of the Osserman conjecture: every globally Osserman manifold, and every pointwise Osserman manifold of dimension $n\ne2,4$, is flat or locally rank-one symmetric. With \cite[Theorem~1]{Nik2010}, part~(2) proves the conformal Osserman conjecture in every dimension $n>4$. \end{remark} The proof starts with the case of two eigenvalues, which also fixes the normalization of $R_{\rm Cay}$. \begin{mainthm}\label{thm:A} Let $R$ be an Osserman algebraic curvature tensor on $\R^{16}$ with exactly two reduced eigenvalues, of multiplicities $7$ and $8$. There are constants $\mu$ and $\kappa\ne0$ such that $\Jac_x=\mu\Pi(x)+\kappa A(x)$, where, in suitable orthogonal coordinates, $A$ has one of the forms \begin{align} A(x)&=\sum_{i=1}^7(J_ix)(J_ix)^T, &&\text{for a $\Cl(7)$-structure},\label{eq:Cl7}\\ A(x)&=\Pi(x)-\sum_{i=1}^8(J_ix)(J_ix)^T, &&\text{for a $\Cl(8)$-structure},\label{eq:Cl8}\\ A(x)&=\frac12\left(\norm{x}^2I+ \sum_{i=0}^8\ip{H_ix}xH_i\right)-xx^T, &&\text{for a symmetric Clifford system $H_0,\ldots,H_8$}. \label{eq:Cayley} \end{align} The first two cases are Clifford. In the third, $R=(\mu-\kappa/3)R_1+(\kappa/3)\,T^*R_{\rm Cay}$ for some $T\in O(16)$. \end{mainthm} \noindent We have not found this classification in the literature: \cite[Lemma~4]{Nik2012} takes it from Conjecture~A, and \cite{AndTwoRoot} treats dimensions twice an odd number only. Corollary~\ref{cor:osserman16}(1) gives the theorem announced by Nikolayevsky in 2006 \cite[Theorem~1.2]{Nik2006}: a pointwise Osserman $16$-manifold with two Jacobi eigenvalues of multiplicities $7$ and $8$ is locally isometric, up to scaling, to the Cayley projective plane or its noncompact dual. Indeed, such a manifold is flat or locally isometric to a rank-one symmetric space, and among these only the Cayley projective plane and its noncompact dual have Jacobi eigenvalues of multiplicities $7$ and $8$: a flat space or a space of constant curvature has one Jacobi eigenvalue, and $\mathbb C\mathrm P^8$, $\mathbb H\mathrm P^4$ and their duals have multiplicities $1,14$ and $3,12$. For an Osserman tensor, \emph{Jacobi-proportionality} means \begin{equation}\label{eq:JP} \norm{x}^2\ip{\pi_i(x)y}y =\norm{y}^2\ip{\pi_i(y)x}x \quad(x,y\ne0,\ \text{every }i), \end{equation} where, if the reduced eigenvalues are $\varepsilon_1\norm x^2,\ldots,\varepsilon_p\norm x^2$, $\pi_i(x)$ is the orthogonal projection onto the $\varepsilon_i\norm x^2$-eigenspace of $\Jac_x$ on $x^\perp$, regarded as an operator on $\R^{16}$ that vanishes on $\R x$; in particular, if $0$ is a reduced eigenvalue, its eigenspace does not contain $x$. \begin{maincor}\label{cor:C} Every Jacobi-proportional Riemannian Osserman algebraic curvature tensor on $\R^{16}$ is Clifford or a linear combination of $R_1$ and an orthogonal conjugate of $R_{\rm Cay}$. \end{maincor} \noindent This settles the multiplicity patterns $(8,7)$ and $(7,7,1)$, to which Andreji\'c reduced the Jacobi-proportional case \cite[Theorem~5]{And2022}, \cite[Theorem~3]{AL2023}. It is proved in Section~\ref{sec:threeroot} from Theorem~\ref{thm:A} and Theorem~\ref{thm:D} below, and it is used in the final step of the proof of Theorem~\ref{thm:conjA}. \begin{mainthm}\label{thm:D} Let $R$ be an Osserman algebraic curvature tensor on $\R^{16}$ with exactly three reduced eigenvalues. If their multiplicities are $(7,7,1)$, $(8,6,1)$, $(8,5,2)$ or $(7,6,2)$, then $R$ is Clifford. \end{mainthm} \paragraph{Organization.} Sections~\ref{sec:quartic}--\ref{sec:Cayley} prove Theorem~\ref{thm:A}: the Jacobi family defines a Cartan--M\"unzner quartic on $\R^{32}$, and the classification of isoparametric hypersurfaces with four principal curvatures of multiplicities $(7,8)$ \cite{Munzner,FKM,Chi} recovers the eigenspaces. Section~\ref{sec:threeroot} proves Theorem~\ref{thm:D} and Corollary~\ref{cor:C} with the polynomial eigenfield method of \cite{Nik2006}. Section~\ref{sec:smallroots} proves Theorem~\ref{thm:E}, which settles fifteen more multiplicity patterns. In the remaining sixteen patterns, Section~\ref{sec:extreme} proves Theorem~\ref{thm:extreme}, the case where the eigenvalue of multiplicity $7$ or $8$ is the smallest or the largest; otherwise we use a rational factorization of the Jacobi family, a coherent factor, whose poles are excluded, or shown to force a Clifford structure, case by case in Sections~\ref{sec:coherent-transpose}--\ref{sec:conjugate-gluing}. Section~\ref{sec:assembly} assembles the cases and proves Theorem~\ref{thm:conjA}. An expanded version of this paper, with many other results, is posted on Zenodo \cite{Gexp}. \section{Normalization and the quartic}\label{sec:quartic} Let $R$ be as in Theorem~\ref{thm:A}. Write the reduced eigenvalues as $\mu\norm{x}^2$ of multiplicity $8$ and $(\mu+\kappa)\norm{x}^2$ of multiplicity $7$, and put $A(x)=(\Jac_x-\mu\Pi(x))/\kappa$. Then $A$ is a homogeneous quadratic symmetric matrix family satisfying \begin{equation}\label{eq:Aproperties} A(x)x=0,\qquad A(x)^2=\norm{x}^2A(x),\qquad \tr A(x)=7\norm{x}^2. \end{equation} The Jacobi family determines its curvature tensor. In fact, for any symmetric quadratic family $A$ with $A(x)x=0$, let $\mathcal B(u,v)=(A(u+v)-A(u)-A(v))/2$. Polarization gives $\mathcal B(x,y)x=-A(x)y/2$. Since $\mathcal B(x,y)$ is symmetric, \[ \ip{A(x)y}y=-2\ip{\mathcal B(x,y)x}y =-2\ip{\mathcal B(y,x)y}x=\ip{A(y)x}x. \] Jacobi-family reconstruction \cite[Theorem 1]{AL2023} therefore applies; specializing its defining formula \cite[(12), p.~8468]{AL2023} to a quadratic family gives the curvature tensor \begin{equation}\label{eq:reconstruct} R_A(u,v)w=\frac23\bigl(\mathcal B(v,w)u-\mathcal B(u,w)v\bigr) \end{equation} with Jacobi family $A$. Conversely, if $R$ is an algebraic curvature tensor with Jacobi family $A$, then $\mathcal B(v,w)u=\tfrac12\bigl(R(u,v)w+R(u,w)v\bigr)$, and the first Bianchi identity turns \eqref{eq:reconstruct} into $R_A=R$; so the Jacobi family determines the tensor. Set \[ Q(x)=A(x)+xx^T,\qquad L(x)=\R x+\operatorname{Im}A(x). \] For $x\ne0$, $Q(x)=\norm{x}^2\pi_{L(x)}$ and $\dim L(x)=8$. In particular, \begin{equation}\label{eq:Qproperties} Q(x)^2=\norm{x}^2Q(x),\qquad Q(x)x=\norm{x}^2x,\qquad \tr Q(x)=8\norm{x}^2. \end{equation} The symmetry $\ip{A(x)y}y=\ip{A(y)x}x$ shown above implies that the biquadratic form \begin{equation}\label{eq:b} b(x,y)=\ip{Q(x)y}y=\ip{A(x)y}y+\ip xy^2 \end{equation} is symmetric in $x,y$. It vanishes when $x=0$ or $y=0$ and satisfies $b(x,x)=\norm{x}^4$. On $\R^{32}=\R^{16}\oplus\R^{16}$ write $z=(x,y)$, $s=\norm{x}^2+\norm{y}^2$, and define \begin{equation}\label{eq:F} F(z)=s^2-8b(x,y). \end{equation} \begin{lemma}\label{lem:quartic} The polynomial $F$ satisfies \[ \norm{\nabla F}^2=16s^3,\qquad \Delta F=8s. \] Its restriction to $S^{31}$ takes both values $1$ and $-1$. \end{lemma} \begin{proof} Symmetry of $b$ gives $\nabla_yb=2Q(x)y$ and $\nabla_xb=2Q(y)x$. By \eqref{eq:Qproperties}, \[ \norm{\nabla b}^2=4sb,\qquad \ip z{\nabla b}=4b,\qquad \Delta b=2\tr Q(x)+2\tr Q(y)=16s. \] Thus \[ \norm{\nabla F}^2 =16s^3-64s\ip z{\nabla b}+64\norm{\nabla b}^2=16s^3. \] Since $\Delta s^2=4(32+2)s=136s$, we obtain $\Delta F=8s$. For unit $x$, $F(x,0)=1$. If $x,y$ are unit and $y\in L(x)$, then $F((x,y)/\sqrt2)=1-8/4=-1$. \end{proof} \Needspace{10\baselineskip} \section{The classification of isoparametric hypersurfaces}\label{sec:external} We record the external results that are used. \paragraph{M\"unzner.} For a homogeneous polynomial of degree $g$ satisfying $\norm{\nabla F}^2=g^2\norm z^{2g-2}$ and $\Delta F=c\norm z^{g-2}$ on $\R^{n+1}$, nonconstant on the sphere, \cite[Satz 3, p.~65]{Munzner} shows that $0$ is a regular value of $F|S^n$ and that the zero level is an isoparametric hypersurface; by its proof (p.~66) the same holds for every level in $(-1,1)$. Its Folgerung (corollary) on p.~66 identifies $F$ with the Cartan polynomial of these hypersurfaces when both $g(n-1)-c$ and $g(n-1)+c$ are nonzero. Satz 4(a), p.~67, gives connected level sets. By Lemma~\ref{lem:quartic} these hypotheses hold with $g=4$, $n=31$ and $c=8$, whereas $g(n-1)=120$. The resulting hypersurfaces have four principal curvatures, and M\"unzner's relations $g(m_1+m_2)=2(n-1)$ and $c=g^2(m_2-m_1)/2$ \cite[Satz~2, p.~63]{Munzner} give \[ m_1+m_2=15,\qquad 8(m_2-m_1)=8, \] so their multiplicities are $(7,8)$. \paragraph{Chi and Ferus--Karcher--M\"unzner.} Chi's classification \cite[abstract, p.~225]{Chi} gives exactly three families with unordered multiplicities $\{7,8\}$. They are the FKM family with parameter $m=7$ and the two FKM families with $m=8$. The first is the Ozeki--Takeuchi family, as identified in \cite[Section 5.9, p.~19]{FKM} (page numbers for FKM refer to its English translation, arXiv:1112.2780); the two $(8,7)$ families are distinguished in its Section 6.6, p.~22. For a symmetric Clifford system $P_0,\ldots,P_m$ on $\R^{32}$ with $m\ge1$, \cite[Theorem 4.1, p.~8; Section 4.3, p.~11]{FKM} gives \begin{equation}\label{eq:FKM} F_P(z)=\norm z^4-2\sum_{i=0}^m\ip{P_iz}z^2,\qquad (m_1,m_2)=(m,15-m). \end{equation} By the same Folgerung of M\"unzner, with $c=\pm8\ne\pm120$, each $F_P$ used here is the Cartan polynomial of its zero level. We need congruence of the polynomials, not only of a level hypersurface. The zero level of $F$ has principal curvatures $\cot(\pi/8+j\pi/4)$, $0\le j\le3$. A congruent member of the FKM family must therefore be its zero member, with one of the two choices of normal. Let $U\in O(32)$ map the zero level of $F$ onto the zero level of $F_P$. Both zero levels are connected \cite[Satz~4(a), p.~67]{Munzner}. By \cite[Folgerung~(iii), p.~66]{Munzner}, $F$ and $F_P\circ U$ are both Cartan polynomials of this hypersurface. The Cartan polynomial of a connected oriented isoparametric hypersurface is determined by it, and reversing the orientation replaces it by its negative \cite[\S4, p.~65]{Munzner}; explicitly $F(rq)=r^4\cos 4t(q)$, where $t(q)\in[0,\pi/4]$ is the spherical distance from $q\in S^{31}$ to the focal submanifold $\{F=1\}\cap S^{31}$ \cite[proof of Satz~3, p.~66; Satz~4(c), p.~67]{Munzner}. Hence $F=\varepsilon F_P\circ U$ with $\varepsilon\in\{1,-1\}$. Absorb $U$ into the matrices $P_i$. Since $\Delta F_P=(136-16(m+1))s$, comparison with $\Delta F=8s$ gives precisely \begin{align} 4b(x,y)&=\sum_{i=0}^7\ip{P_iz}z^2, \label{eq:plus}\\ \text{or}\qquad 4b(x,y)&=s^2-\sum_{i=0}^8\ip{P_iz}z^2. \label{eq:minus} \end{align} These identities hold in the original coordinates $(x,y)$; no preservation of the two coordinate summands by $U$ has been assumed. \section{A lemma on quadratic maps between spheres} Quadratic maps between spheres have a classical literature \cite{Wood,Shapiro}. We give the short argument needed here. Part~(b) below is the quadratic case of \cite[p.~163]{Wood}; we include its short proof. \begin{lemma}\label{lem:threshold} Let $H_0,\ldots,H_m$ be symmetric $n\times n$ matrices and $h(x)=(\ip{H_ix}x)_{i=0}^m$. Suppose $\norm{h(x)}=\norm x^2$. For $a\in\R^{m+1}$ put $H(a)=\sum_i a_iH_i$. \begin{enumerate}[label=\textup{(\alph*)}] \item If $a\in h(S^{n-1})$, then $I-H(a)\ge0$ and the fibre over $a$ is the unit sphere of $U_a=\ker(I-H(a))$, where $\dim U_a\ge n-m$. \item If $n>2m$, then $h$ is constant on $S^{n-1}$. \item If $n=2m$, $m\ge1$, and $h$ is not constant on $S^{n-1}$, then $H_iH_j+H_jH_i=2\delta_{ij}I$. \end{enumerate} \end{lemma} \begin{proof} For unit $a,x$, $\ip{H(a)x}x=\ip a{h(x)}\le1$, with equality exactly when $h(x)=a$. This proves the fibre description. Fix a unit $x$ in that fibre and rotate the index space so that $a=e_0$. Then $H_0x=x$. Comparing the $t^2$ coefficients of $\norm{h(x+ty)}^2=\norm{x+ty}^4$ gives \[ 2\ip{H_0y}y+4\sum_{i=0}^m\ip{H_ix}y^2 =2\norm y^2+4\ip xy^2. \] The $i=0$ terms cancel, so \begin{equation}\label{eq:rankidentity} I-H_0=2\sum_{i=1}^m(H_ix)(H_ix)^T. \end{equation} Its rank is at most $m$, proving (a). The subspaces $U_a,U_b$ for distinct values $a\ne b$ meet only at zero. If $n>2m$, two such subspaces would have dimensions summing to more than $n$. This proves (b). Suppose $n=2m$ and there is more than one value. For every $a$, comparison with another fibre gives $\dim U_a=m$. Hence the right side of \eqref{eq:rankidentity} has rank $m$, and $H_1x,\ldots,H_mx$ are independent and perpendicular to $x$. On $T_xS^{n-1}=x^\perp$, \[ (dh)_x(y)=2(\ip{H_ix}y)_{i=0}^m \] has rank $m$. The image of $h$ is open in $S^m$ and is compact; since $m\ge1$, it is all of $S^m$. In particular $a$ and $-a$ are values. Their fibres are the $+1$ and $-1$ eigenspaces of $H(a)$ and each has dimension $m$. Thus $H(a)^2=I$ for every unit $a$. Polarization in $a$ proves (c). \end{proof} \section{Recovering the original eigenspaces} \begin{proposition}\label{prop:plus} Identity \eqref{eq:plus} implies \eqref{eq:Cl7}. \end{proposition} \begin{proof} Setting $y=0$ in \eqref{eq:plus} shows that the upper-left diagonal block of each $P_i$ is zero; setting $x=0$ gives the same for the lower-right block. Hence \[ P_i=\begin{pmatrix}0&C_i^T\\ C_i&0\end{pmatrix}, \qquad Q(x)=\sum_{i=0}^7(C_ix)(C_ix)^T . \] The Clifford relations give \[ C_i^TC_j+C_j^TC_i=2\delta_{ij}I,\qquad C_iC_j^T+C_jC_i^T=2\delta_{ij}I. \] Moreover $Q(x)x=\norm x^2x$ implies $\sum_i\ip{C_ix}x^2=\norm x^4$. Apply Lemma~\ref{lem:threshold}(b) to the symmetric parts of $C_i$, with $n=16$ and $m=7$. There is a unit $w\in\R^8$ such that $(\ip{C_ix}x)_i=\norm x^2w$. Rotating the index space gives $w=e_0$. Since $C_0$ is orthogonal and $\ip{C_0x}x=\norm x^2$, we have $C_0=I$. The remaining $C_i$ are skew, and their relations become $C_iC_j+C_jC_i=-2\delta_{ij}I$. Subtracting $xx^T$ from $Q(x)$ proves the claim. \end{proof} \begin{proposition}\label{prop:minus} Identity \eqref{eq:minus} implies \eqref{eq:Cl8} or \eqref{eq:Cayley}. \end{proposition} \begin{proof} In the original splitting write $P_i=\left(\begin{smallmatrix}H_i&B_i^T\\ B_i&D_i\end{smallmatrix}\right)$. Setting $y=0$ and then $x=0$ in \eqref{eq:minus} gives \[ \sum_{i=0}^8\ip{H_ix}x^2=\norm x^4,\qquad \sum_{i=0}^8\ip{D_iy}y^2=\norm y^4. \] Write $h(x)=(\ip{H_ix}x)_i$ and $d(y)=(\ip{D_iy}y)_i$. If $h$ is constant on the unit sphere, rotate the index so that $H_0=I$, $H_i=0$ for $i\ge1$. The upper block of $P_0^2=I$ gives $B_0=0$. Also $\tr P_0=0$, because $P_0=-P_1P_0P_1$. Thus $D_0^2=I$ and $\tr D_0=-16$, forcing $D_0=-I$. Anticommutation with $P_0=\operatorname{diag}(I,-I)$ makes the other eight matrices off-diagonal. Calling their lower blocks $C_1,\ldots,C_8$ and expanding \eqref{eq:minus} yields \[ Q(x)=\norm x^2I-\sum_{i=1}^8(C_ix)(C_ix)^T. \] The identity $Q(x)x=\norm x^2x$ gives $\sum_i\ip{C_ix}x^2=0$. Each $C_i$ is therefore skew, and the Clifford relations become those of a $\Cl(8)$-structure. This is \eqref{eq:Cl8}. If $h$ is not constant, Lemma~\ref{lem:threshold}(c) makes the $H_i$ a symmetric Clifford system. The equation $H_i^2+B_i^TB_i=I$ forces $B_i=0$. Expanding \eqref{eq:minus} now gives \[ b(x,y)=\frac12\left(\norm x^2\norm y^2-\ip{h(x)}{d(y)}\right). \] Since $b(x,x)=\norm x^4$ and $\norm{h(x)}=\norm{d(x)}=\norm x^2$, equality in Cauchy--Schwarz forces $d(x)=-h(x)$. Polarization gives $D_i=-H_i$. Consequently \[ Q(x)=\frac12\left(\norm x^2I+\sum_{i=0}^8\ip{H_ix}xH_i\right), \] which proves \eqref{eq:Cayley}. \end{proof} \section{Identification of the Cayley tensor}\label{sec:Cayley} We first prove two identities for an arbitrary symmetric Clifford system $S_0,\ldots,S_8$ on $\R^{16}$: \eqref{eq:HopfM} below and the norm identity $\sum_{i=0}^8\ip{S_ix}x^2=\norm x^4$. For $q=\norm x^2$ put $h_i=\ip{S_ix}x$ and $S_h=\sum_i h_iS_i$. The products of zero, one, and four distinct generators are symmetric (a product of $k$ distinct generators is $(-1)^{k(k-1)/2}$ times its transpose) and trace-orthogonal (the product of two distinct such monomials is, up to sign, a product of $j$ distinct generators with $1\le j\le8$, which anticommutes with a generator outside it if $j$ is odd and with one inside it if $j$ is even, so has trace zero), and their number is $1+9+126=136$. They therefore form a basis of the symmetric matrices. If $\mathcal L_k$ is their span for degree $k$, the operator $\mathcal T(Y)=\sum_i S_iYS_i$ acts on $\mathcal L_k$ by $(-1)^k(9-2k)$, namely by $9,-7,1$ for $k=0,1,4$. These assertions also appear in \cite[Lemma~3, pp.~6--7 of arXiv:0910.1645v1]{Nik2012}. Expanding $xx^T=qI/16+S_h/16+Y_4$ therefore gives \begin{equation}\label{eq:HopfM} Z_x:=\sum_i(S_ix)(S_ix)^T =xx^T+\frac12(qI-S_h). \end{equation} Contraction with $x$ yields $\norm h^2=q^2+(q^2-\norm h^2)/2$, so $\norm h=q$. For this system define $Q(x)=(qI+S_h)/2$, so that $A(x)=Q(x)-xx^T$ is the matrix in \eqref{eq:Cayley} with $H_i=S_i$. For $x\ne0$ set $w=h/q$. Since $S_h=qS_w$, we have $Q(x)=q(I+S_w)/2$. Also $S_wx=x$, since $S_w$ is an orthogonal involution and $\ip{S_wx}x=q$. Its two eigenspaces have dimension eight. For $u\perp w$, $S_ux$ lies in its negative eigenspace. With \eqref{eq:HopfM} and $S_h=qS_w$, \[ Q(x)=\frac q2(I+S_w),\qquad Z_x=xx^T+qI-Q(x),\qquad S_h=2Q(x)-qI . \] In octonionic coordinates $x=(x_1,x_2)$ the standard system is \[ S_0(x_1,x_2)=(x_1,-x_2),\qquad S_i(x_1,x_2)=(e_i\overline{x_2},\,\overline{x_1}e_i),\quad1\le i\le8, \] where $(e_i)_{i=1}^8$ is an orthonormal basis of $\OCT$. With $(u\wedge v)z=\ip uz v-\ip vz u$, the Brown--Gray Cayley curvature formula \cite[Eq.~(6.12)]{BG}, in the form \cite[equation~(7), p.~6 of arXiv:0910.1645v1]{Nik2012}, is \[ R_{\OCT}(u,v)=3u\wedge v+\sum_{i=0}^8 S_iu\wedge S_iv. \] In our Jacobi convention this gives \begin{align*} \Jac_x(R_{\OCT}) &=3(xx^T-qI)+Z_x-S_h\\ &=4xx^T-qI-3Q(x)=-\Pi(x)-3A(x). \end{align*} Therefore $R_{\rm Cay}=-R_{\OCT}$ has Jacobi family $\Pi(x)+3A(x)$, with reduced eigenvalues $1$ and $4$ of multiplicities eight and seven. Let $\Cl_+(9)$ denote the real Clifford algebra on nine generators with squares $+1$ and pairwise anticommutation. Then $\Cl_+(9)\cong M_{16}(\R)\oplus M_{16}(\R)$. Its representations on $\R^{16}$ are equivalent up to orthogonal conjugacy and the sign of one generator; see \cite[Sections 3.4--3.5, pp.~6--7]{FKM} and \cite[p.~6 of arXiv:0910.1645v1]{Nik2012}. Changing a generator's sign leaves $\ip{S_ix}xS_i$ unchanged. This identifies every family \eqref{eq:Cayley} with the standard Cayley family, so $3A(x)=\Jac_x(T^*R_{\rm Cay})-\Pi(x)$ for some $T\in O(16)$ and $\Jac_x=(\mu-\kappa/3)\Pi(x)+(\kappa/3)\Jac_x(T^*R_{\rm Cay})$. As the Jacobi family determines the curvature tensor by \eqref{eq:reconstruct}, $R=(\mu-\kappa/3)R_1+(\kappa/3)\,T^*R_{\rm Cay}$. This finishes the proof of Theorem~\ref{thm:A}: by Lemma~\ref{lem:quartic} and Section~\ref{sec:external}, $b$ satisfies \eqref{eq:plus} or \eqref{eq:minus}; Propositions~\ref{prop:plus} and~\ref{prop:minus} give the forms \eqref{eq:Cl7}--\eqref{eq:Cayley}, of which the first two are Clifford and the third has just been identified. \section{Tensors with three eigenvalues, and Jacobi-proportionality}\label{sec:threeroot} For the standard Cayley family of Section~\ref{sec:Cayley}, that is \eqref{eq:Cayley} with $H_i=S_i$, let $E(x)=L(x)\cap x^\perp$ and $N(x)=L(x)^\perp$, of dimensions seven and eight; here $L(x)=\R x+\operatorname{Im}A(x)$ as in Section~\ref{sec:quartic}. \begin{lemma}\label{lem:nofield} There is no nonzero linear map $K$ with $Kx\in E(x)$ for every $x\ne0$. There is also no nonzero linear map $K$ with $Kx\in N(x)$ for every $x\ne0$. \end{lemma} \begin{proof} After the orthogonal coordinate change $(x_1,x_2)\mapsto(x_1,\overline{x_2})$, the Cayley lines are \[ \mathcal C_\omega=\{(u,\omega u):u\in\OCT\},\quad \omega\in\OCT,\qquad \mathcal C_\infty=\{(0,v):v\in\OCT\}. \] Here $S_{(w_0,a)}=w_0S_0+\sum_{i=1}^8a_iS_i$ for $w_0^2+\norm a^2=1$ and $a=\sum_i a_i e_i$. Its positive eigenspace becomes $\mathcal C_\omega$ with $\omega=\overline a/(1+w_0)$ when $w_0\ne-1$, and becomes $\mathcal C_\infty$ otherwise. Distinct lines among the $\mathcal C_\omega$ and $\mathcal C_\infty$ meet only in $0$, as $\OCT$ has no zero divisors, and $x\in L(x)$; so for $x$ in such a line, $L(x)$ is that line. Suppose first $Kx\in E(x)$. Since $\ip{Kx}x=0$, $K$ is skew. Preservation of $\mathcal C_0,\mathcal C_\infty,\mathcal C_1$ forces $K=\operatorname{diag}(D,D)$. Preservation of $\mathcal C_\omega$ gives $D(\omega u)=\omega D(u)$. Put $c=D(1)$; then $D(u)=uc$ and $(\omega u)c=\omega(uc)$ for all $\omega,u$. Thus $c$ lies in the right nucleus of $\OCT$, which is $\R$. Hence $K=cI$, and skewness forces $c=0$. Suppose next $Kx\in N(x)$. The lines $\mathcal C_0$ and $\mathcal C_\infty$ force $K=\left(\begin{smallmatrix}0&B\\ C&0\end{smallmatrix}\right)$. The graph condition for $\mathcal C_\omega^\perp$ is $B(\omega u)=-\overline\omega C(u)$. Taking $\omega=1$ gives $B=-C$; taking $u=1$ then gives $C(\omega)=\overline\omega\gamma$, where $\gamma=C(1)$. Substitution yields \[ (ab)\gamma=b(a\gamma)\qquad(a,b\in\OCT). \] Choose a quaternion subalgebra containing $\gamma$ and choose noncommuting $a,b$ in it. Associativity in that subalgebra gives $(ab-ba)\gamma=0$. Multiplicativity of the octonion norm forces $\gamma=0$, and therefore $K=0$. \end{proof} \begin{proof}[Proof of Theorem~\ref{thm:D}] Let $t\in\{1,2\}$ be the smallest multiplicity, with eigenvalue $\tau$. To apply \cite[Lemma~3.1]{Nik2006}, shift a different reduced eigenvalue to zero by subtracting a multiple of $R_1$. In the convention of \cite[p.~379]{Nik2006}, the shifted tensor has $k=3$ reduced eigenvalues, one of them zero. Nikolayevsky's \cite[Lemma 3.1(1),(2), p.~383]{Nik2006} gives an orthonormal frame for the $\tau$-eigenspace by odd homogeneous polynomial fields of degree at most $k-1=2$. They are therefore linear: $J_1x,\ldots,J_tx$. Their norm and orthogonality identities make each $J_i$ skew and orthogonal and make the $J_i$ anticommute. Return to the original tensor, whose eigenspaces are unchanged by the shift. Choose another reduced eigenvalue $\tau'$ so that moving $\tau$ to $\tau'$ merges the multiplicities as \[ (7,7,1)\to(7,8),\quad (8,6,1)\to(8,7),\quad (8,5,2)\to(8,7),\quad (7,6,2)\to(7,8). \] For a skew orthogonal $J$, let $R_J$ denote the curvature tensor whose Jacobi family is $(Jx)(Jx)^T$, obtained for example from \eqref{eq:reconstruct}. Then \[ \widetilde R=R-(\tau-\tau')\sum_{i=1}^t R_{J_i} \] is Osserman with two distinct eigenvalues of multiplicities seven and eight. Apply Theorem~\ref{thm:A}. Its Cayley alternative is impossible: each $J_ix$ remains an eigenfield in one of the two merged distributions, contradicting Lemma~\ref{lem:nofield} after conjugating the field to the standard model: $K=TJ_iT^{-1}$. Thus $\widetilde R$ is Clifford. Restore the rank-one terms successively. By \cite[Lemma 3.2(1), pp.~383--384]{Nik2006}, if $R$ is Osserman, $J$ is orthogonal and skew, $c\ne0$, and $R-cR_J$ is Osserman and Clifford, then $R$ is Clifford. Each intermediate tensor here is Osserman: only the constant eigenvalues on the mutually orthogonal eigenlines $\R J_ix$ change. The lemma consequently applies at every step, in the form of Remark~\ref{rem:lemma32-cl8} whenever $R-cR_J$ has a Clifford structure with $\nu=8$ (an endpoint of type \eqref{eq:Cl8}, or the intermediate pattern $(7,7,1)$ reached from $(7,6,2)$). The sign convention of \cite{Nik2006} is opposite to ours, which changes the sign of $c$; only $c\ne0$ is used. This proves the theorem. \end{proof} \begin{remark}\label{rem:lemma32-cl8} We recall the proof of \cite[Lemma~3.2(1), pp.~383--384]{Nik2006} in the form used here. The proof in \cite{Nik2006} shifts the eigenvalue of largest multiplicity of $R-cR_J$ to zero; the version below shifts instead the base constant $\lambda_0$ of a Clifford structure of $R-cR_J$. The two choices differ only when $\nu=8$ and another eigenvalue also has multiplicity at least seven, that is, at an endpoint of type \eqref{eq:Cl8} or at the pattern $(7,7,1)$. Indeed, with the generators of zero coefficient dropped, $\lambda_0$ has multiplicity $15-\nu$; for $\nu\le7$ it is the unique eigenvalue of largest multiplicity, for $\nu=8$ it has multiplicity seven, and at $(7,7,1)$ every Clifford structure has $\nu=8$. Let $\Jac_x(R-cR_J)=\lambda_0\Pi(x)+\sum_{\alpha=1}^{\nu}\mu_\alpha(J_\alpha x)(J_\alpha x)^T$ with $\mu_\alpha\ne0$; subtracting $\lambda_0R_1$ from $R$ we may take $\lambda_0=0$. Let $\lambda_1,\ldots,\lambda_p$ be the distinct values of the $\mu_\alpha$, and $\lambda_0=0$. For unit $x$ the eigenspaces of $\Jac_x(R-cR_J)$ on $x^\perp$ are $E_i(x)=\Span\{J_\alpha x:\mu_\alpha=\lambda_i\}$ for $i\ge1$ and $E_0(x)=x^\perp\cap\bigl(\sum_\alpha\R J_\alpha x\bigr)^\perp$; let $\pi_i$ be the orthogonal projections onto them. As $\Jac_x(R)=\Jac_x(R-cR_J)+c(Jx)(Jx)^T$ and $Jx\perp x$, \[ \frac{\det\bigl(tI-\Jac_x(R)\bigr)}{\det\bigl(tI-\Jac_x(R-cR_J)\bigr)} =1-c\sum_{i=0}^{p}\frac{\norm{\pi_i(Jx)}^2}{t-\lambda_i} \] on $x^\perp$, so, both tensors being Osserman, partial fractions make each $\norm{\pi_i(Jx)}=c_i$ constant on the unit sphere. For $i\ge1$, $\sum_{\mu_\alpha=\lambda_i}\ip{J_\alpha x}{Jx}^2=c_i^2\norm x^4$ is a sum of at most eight squares divisible by $\norm x^2$, so by \cite[Lemma~3.11]{Nik2006} each $\ip{J_\alpha x}{Jx}$ (a quadratic form) is a constant multiple of $\norm x^2$; after an orthogonal change of the $J_\alpha$ with $\mu_\alpha=\lambda_i$ we get $\pi_i(Jx)=c_iJ^{(i)}x$ for one of them, $J^{(i)}$. Then $J_{\nu+1}x:=\pi_0(Jx)=Jx-\sum_{i\ge1}c_iJ^{(i)}x$ is linear and skew, anticommutes with every $J_\alpha$ (as $J_{\nu+1}x\perp J_\alpha x$), and $\norm{J_{\nu+1}x}^2=c_0^2\norm x^2$ with $c_0^2=1-\sum_{i\ge1}c_i^2$. If $\nu=8$, then $c_0=0$: otherwise $c_0^{-1}J_9$ would be a ninth orthogonal complex structure on $\R^{16}$ anticommuting with $J_1,\ldots,J_8$, and $\R^{16}$ carries at most eight (Hurwitz--Radon). Finally, write $Jx=\sum_{i\ge1}c_iJ^{(i)}x+c_0\bigl(c_0^{-1}J_{\nu+1}x\bigr)$, omitting the last term if $c_0=0$. An orthogonal change of $J^{(1)},\ldots,J^{(p)}$ (and of $c_0^{-1}J_{\nu+1}$ if $c_0\ne0$) that diagonalizes $\operatorname{diag}(\lambda_1,\ldots,\lambda_p,0)+c\,\gamma\gamma^T$, $\gamma=(c_1,\ldots,c_p,c_0)^T$ (without the last row and column if $c_0=0$), gives a Clifford structure for $R$. \end{remark} \begin{proof}[Proof of Corollary~\ref{cor:C}] Andreji\'c \cite[Theorem 5]{And2022}, restated as \cite[Theorem 3, p.~8470]{AL2023}, proves that a Riemannian Jacobi-proportional Osserman algebraic curvature tensor on $\R^{16}$ that is not Clifford has multiplicity pattern $(8,7)$ or $(7,7,1)$. Andreji\'c's hypothesis places no restriction on the number of eigenvalues. Apply Theorems~\ref{thm:A} and \ref{thm:D}. \end{proof} \input{sec_small_roots} \input{sec_polynomial_gram} \input{sec_coherent_transpose_poles} \input{sec_real_pole_types} \input{sec_common_lemmas} \input{sec_rank_seven_no_poles} \input{sec_simple_rank_four_pole} \input{sec_double_quadratic_exclusion} \input{sec_closed_kernel_and_ten_space} \input{sec_small_support_incidence} \input{sec_aligned_quadratic_two_poles} \input{sec_quadratic_proper_supports} \input{sec_real_rank_four_images} \input{sec_divisor_budget} \input{sec_balanced_alignment} \input{sec_positive_model_factorial} \input{sec_singular_integral_extraction} \input{sec_conjugate_preliminary_reductions} \input{sec_conjugate_gluing} \input{sec_final_assembly} \section*{Acknowledgments}\addcontentsline{toc}{section}{Acknowledgments} This work was done overwhelmingly by AI models and supervised by the author. The models were Anthropic's Claude and OpenAI's ChatGPT. The author takes full responsibility for the paper. \section*{Statements and Declarations}% \addcontentsline{toc}{section}{Statements and Declarations} \noindent\textbf{Funding.} No funding was received for conducting this study. \smallskip\noindent\textbf{Competing interests.} The author has no competing interests to declare that are relevant to the content of this article. \smallskip\noindent\textbf{Data availability.} No datasets were generated or analysed during the current study. The Lean files are available at the GitHub link in the footnote on the first page; the expanded version is \cite{Gexp}. \phantomsection\addcontentsline{toc}{section}{References} \begin{thebibliography}{99} \bibitem{And2022} V.~Andreji\'c, \emph{The proportionality principle for Osserman manifolds}, J. Geom. Phys. \textbf{176} (2022), 104516. \href{https://doi.org/10.1016/j.geomphys.2022.104516}{doi:10.1016/j.geomphys.2022.104516}. \bibitem{AndTwoRoot} V.~Andreji\'c, \emph{Two-root Riemannian manifolds}, Mediterr. J. Math. \textbf{20} (2023), article 100. \href{https://doi.org/10.1007/s00009-023-02340-2}{doi:10.1007/s00009-023-02340-2}; \href{https://arxiv.org/abs/2009.12834}{arXiv:2009.12834v4}. \bibitem{AL2023} V.~Andreji\'c and K.~Luki\'c, \emph{On the existence of a curvature tensor for given Jacobi operators}, Filomat \textbf{37}:25 (2023), 8465--8471. \href{https://doi.org/10.2298/FIL2325465A}{doi:10.2298/FIL2325465A}. \url{https://www.pmf.ni.ac.rs/filomat-content/2023/37-25/37-25-6-19416.pdf}. \bibitem{BCR1998} J.~Bochnak, M.~Coste and M.-F.~Roy, \emph{Real Algebraic Geometry}, Ergeb. Math. Grenzgeb. (3) \textbf{36}, Springer, Berlin, 1998. \href{https://doi.org/10.1007/978-3-662-03718-8}{doi:10.1007/978-3-662-03718-8}. \bibitem{BochnakKucharz} J.~Bochnak and W.~Kucharz, \emph{On homology classes represented by real algebraic varieties}, Banach Center Publ. \textbf{44} (1998), 21--35. \url{https://eudml.org/doc/208886}. \bibitem{BG} R.~B. Brown and A.~Gray, \emph{Riemannian manifolds with holonomy group $\operatorname{Spin}(9)$}, in \emph{Differential Geometry (in Honor of Kentaro Yano)}, Kinokuniya, Tokyo, 1972, 41--59. \bibitem{Calabi1964} E.~Calabi, \emph{Linear systems of real quadratic forms}, Proc. Amer. Math. Soc. \textbf{15} (1964), 844--846. \href{https://doi.org/10.1090/S0002-9939-1964-0166203-7}{doi:10.1090/S0002-9939-1964-0166203-7}. \bibitem{Chi1988} Q.-S. Chi, \emph{A curvature characterization of certain locally rank-one symmetric spaces}, J. Differential Geom. \textbf{28} (1988), 187--202. \href{https://doi.org/10.4310/jdg/1214442277}{doi:10.4310/jdg/1214442277}. \bibitem{Chi} Q.-S. Chi, \emph{Isoparametric hypersurfaces with four principal curvatures, IV}, J. Differential Geom. \textbf{115} (2020), 225--301. \href{https://doi.org/10.4310/JDG/1589853626}{doi:10.4310/JDG/1589853626}; \href{https://arxiv.org/abs/1605.00976}{arXiv:1605.00976}. \bibitem{EGHP2004} D.~Eisenbud, M.~Green, K.~Hulek and S.~Popescu, \emph{Small schemes and varieties of minimal degree}, Amer. J. Math. \textbf{128} (2006), 1363--1389. \href{https://doi.org/10.1353/ajm.2006.0043}{doi:10.1353/ajm.2006.0043}; page numbers refer to \href{https://arxiv.org/abs/math/0404517}{arXiv:math/0404517}. \bibitem{FKM} D.~Ferus, H.~Karcher and H.-F. M\"unzner, \emph{Cliffordalgebren und neue isoparametrische Hyperfl\"achen}, Math. Z. \textbf{177} (1981), 479--502. \href{https://doi.org/10.1007/BF01219082}{doi:10.1007/BF01219082}. English translation by T.~E. Cecil, \emph{Clifford algebras and new isoparametric hypersurfaces}, \href{https://arxiv.org/abs/1112.2780}{arXiv:1112.2780}. \bibitem{Finsler1937} P.~Finsler, \emph{\"Uber das Vorkommen definiter und semidefiniter Formen in Scharen quadratischer Formen}, Comment. Math. Helv. \textbf{9} (1937), 188--192. \href{https://doi.org/10.1007/BF01258188}{doi:10.1007/BF01258188}. \bibitem{GKV2002} E.~Garc\'ia-R\'io, D.~N. Kupeli and R.~V\'azquez-Lorenzo, \emph{Osserman Manifolds in Semi-Riemannian Geometry}, Lecture Notes in Math. \textbf{1777}, Springer, Berlin, 2002. \href{https://doi.org/10.1007/b83213}{doi:10.1007/b83213}. \bibitem{Gilkey1994} P.~B. Gilkey, \emph{Manifolds whose curvature operator has constant eigenvalues at the basepoint}, J. Geom. Anal. \textbf{4} (1994), no.~2, 155--158. \href{https://doi.org/10.1007/BF02921544}{doi:10.1007/BF02921544}. \bibitem{Gilkey2001} P.~B. Gilkey, \emph{Geometric Properties of Natural Operators Defined by the Riemann Curvature Tensor}, World Scientific, River Edge, NJ, 2001. \bibitem{GSV1995} P.~Gilkey, A.~Swann and L.~Vanhecke, \emph{Isoparametric geodesic spheres and a conjecture of Osserman concerning the Jacobi operator}, Quart. J. Math. Oxford Ser.~(2) \textbf{46} (1995), no.~3, 299--320. \href{https://doi.org/10.1093/qmath/46.3.299}{doi:10.1093/qmath/46.3.299}. \bibitem{Gexp} D.~Goel, \emph{The Osserman (1990) conjecture is fully resolved, and other miscellanea}, expanded version, Zenodo, 2026. \href{https://doi.org/10.5281/zenodo.23603278}{doi:10.5281/zenodo.23603278}. \bibitem{EGAIV3} A.~Grothendieck, \emph{\'El\'ements de g\'eom\'etrie alg\'ebrique. IV. \'Etude locale des sch\'emas et des morphismes de sch\'emas, troisi\`eme partie}, Publ. Math. IH\'ES \textbf{28} (1966), 5--255. \href{https://doi.org/10.1007/BF02684343}{doi:10.1007/BF02684343}. \bibitem{Kryszewski} W.~Kryszewski, \emph{Remarks to the Vietoris theorem}, Topol. Methods Nonlinear Anal. \textbf{8} (1996), 383--405. \href{https://doi.org/10.12775/TMNA.1996.041}{doi:10.12775/TMNA.1996.041}; \url{https://projecteuclid.org/journals/topological-methods-in-nonlinear-analysis/volume-8/issue-2/Remarks-to-the-Vietoris-theorem/tmna/1479265248.full}. \bibitem{MartinezLegaz2005} J.~E.~Mart\'inez-Legaz, \emph{On Brickman's theorem}, J. Convex Anal. \textbf{12} (2005), 139--143. \url{https://www.heldermann-verlag.de/jca/jca12/jca0474_b.pdf}. \bibitem{Munzner} H.-F. M\"unzner, \emph{Isoparametrische Hyperfl\"achen in Sph\"aren. I}, Math. Ann. \textbf{251} (1980), 57--71. \href{https://doi.org/10.1007/BF01420281}{doi:10.1007/BF01420281}. \bibitem{Nik2004} Y.~Nikolayevsky, \emph{Osserman manifolds of dimension 8}, Manuscripta Math. \textbf{115} (2004), 31--53. \href{https://doi.org/10.1007/s00229-004-0480-y}{doi:10.1007/s00229-004-0480-y}; \href{https://arxiv.org/abs/math/0310387}{arXiv:math/0310387}. \bibitem{Nik2005} Y.~Nikolayevsky, \emph{Osserman conjecture in dimension $n\ne8,16$}, Math. Ann. \textbf{331} (2005), 505--522. \href{https://doi.org/10.1007/s00208-004-0580-8}{doi:10.1007/s00208-004-0580-8}; \href{https://arxiv.org/abs/math/0204258v2}{arXiv:math/0204258v2}; page numbers cited from the arXiv version refer to v2. \bibitem{Nik2006} Y.~Nikolayevsky, \emph{On Osserman manifolds of dimension 16}, in \emph{Contemporary Geometry and Related Topics} (N.~Bokan et al., eds.), Proceedings of the Conference, Belgrade, June 26--July 2, 2005, Faculty of Mathematics, University of Belgrade, 2006, 379--398. \url{https://www.maths.tcd.ie/EMIS/proceedings/CGRT2005/Articles/cgrt24.pdf}. \bibitem{Nik2010} Y.~Nikolayevsky, \emph{Conformally Osserman manifolds}, Pacific J. Math. \textbf{245} (2010), 315--358. \href{https://doi.org/10.2140/pjm.2010.245.315}{doi:10.2140/pjm.2010.245.315}; \href{https://arxiv.org/abs/0810.5621}{arXiv:0810.5621}. \bibitem{Nik2012} Y.~Nikolayevsky, \emph{Conformally Osserman manifolds of dimension 16 and a Weyl--Schouten theorem for rank-one symmetric spaces}, Ann. Mat. Pura Appl. \textbf{191} (2012), 677--709 (published online 2011). \href{https://doi.org/10.1007/s10231-011-0201-6}{doi:10.1007/s10231-011-0201-6}. Preprint: \emph{Osserman manifolds and the Weyl--Schouten theorem for rank-one symmetric spaces}, \href{https://arxiv.org/abs/0910.1645v1}{arXiv:0910.1645v1}. \bibitem{NR2013} Y.~Nikolayevsky and Z.~Raki\'c, \emph{A note on Raki\'c duality principle for Osserman manifolds}, Publ. Inst. Math. (Beograd) (N.S.) \textbf{94}(108) (2013), 43--45. \href{https://doi.org/10.2298/PIM1308043N}{doi:10.2298/PIM1308043N}; \href{https://arxiv.org/abs/1204.1600}{arXiv:1204.1600}. \bibitem{Oss1990} R.~Osserman, \emph{Curvature in the eighties}, Amer. Math. Monthly \textbf{97} (1990), 731--756. \href{https://doi.org/10.1080/00029890.1990.11995659}{doi:10.1080/00029890.1990.11995659}. \bibitem{Samuel1964} P.~Samuel, \emph{Lectures on Unique Factorization Domains}, Tata Institute of Fundamental Research, Bombay, 1964. \url{https://mathweb.tifr.res.in/Documents/Publications/Lectures/tifr30.pdf}. \bibitem{Shapiro} D.~B. Shapiro, \emph{Compositions of Quadratic Forms}, de Gruyter Expositions in Mathematics \textbf{33}, Walter de Gruyter, Berlin--New York, 2000. \bibitem{Wood} R.~Wood, \emph{Polynomial maps from spheres to spheres}, Invent. Math. \textbf{5} (1968), 163--168. \href{https://doi.org/10.1007/BF01425547}{doi:10.1007/BF01425547}. \end{thebibliography} \end{document}