How statement and proof provenance work
The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.
- Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
- AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
- AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.
These labels describe origin, not correctness: citations and verification chips remain separate evidence.
Exclusions for positive definite diagrams: trees, valency, labels, chains and arms
Statement
Let be a finite set with Coxeter matrix and diagram (Coxeter diagrams: edges, labels, components and finite type), let and let be the Coxeter form (The real Coxeter form, its radical, reflections, and form-preserving maps). For in write so that , exactly when , and whenever . The cosine matrix is , so has diagonal entries and off-diagonal entries . Assume that is connected and that is positive definite (Positive and negative definiteness, the inertia , rank , and signature of a real symmetric bilinear or quadratic form). Then:
(1) Witness principle. If and there is with , then is not positive definite, because such a is a nonzero vector of with . Moreover, if all coordinates of in the basis are and a labelled graph on has all labels at most the corresponding labels of (a non-edge having label ), then , where is the cosine form of ; so a non-positive value of on a non-negative vector excludes positive definiteness of . The explicit non-negative witnesses below use this comparison.
(2) No cycles. contains no cycle: if () are distinct vertices whose consecutive pairs ( modulo ) are edges, then satisfies , because the consecutive pairs contribute each and all other pairs contribute .
(3) Valency and local labels. For every , In particular: no label is ; no vertex has four or more neighbours; if a vertex has exactly three neighbours then its three edges all have label ; and if a vertex has exactly two neighbours with labels , then and ; in particular the pairs and are forbidden at a vertex.
(4) At most one branch vertex, and at most one large label.
(i) has at most one vertex of degree (hence, with (3), at most one vertex of degree ).
(ii) has at most one edge whose label is ; and if such an edge exists then has no vertex of degree , so by (3) is a path.
(5) Paths. Suppose is a path on vertices, with labels along the path. In formulas involving labels, denotes the coefficient .
(i) If is the determinant of the leading principal submatrix of the cosine matrix , then , and for ; for the path with all labels one gets for every , and .
(ii) If the labels are all except one edge labelled , and that edge splits the path into two subpaths with and vertices (, , ), then Consequently: if then ; if then ; if then , or .
(6) Three arms. Suppose has a (unique) vertex of degree and all its edges have label , and let be the numbers of vertices in the three components of (each of which is a path). Then Consequently, up to permutation, for some , or .
(7) Conclusion. Every connected positive definite Coxeter diagram is isomorphic as a labelled graph to one of: (; a path, all labels ), (; a path, labels ), (; the star with arms of vertices), (the stars with arms ; ; ), (the path with labels ), (the path with labels ), (the path with labels ), or (; two vertices joined by one edge labelled ).
Facts & Assumptions
Given: A finite set with Coxeter matrix , its diagram , the space with the Coxeter form , the cosine numbers of the statement, and the hypothesis that is connected and is positive definite.
In the diagram , distinct vertices are joined by an edge exactly when , and the neighbours of are ; the components of partition , and a cycle of is a cycle of the underlying simple graph (Coxeter diagrams: edges, labels, components and finite type).
is the unique symmetric bilinear form on with , for finite and for ; in particular is a symmetric matrix and (The real Coxeter form, its radical, reflections, and form-preserving maps, Bilinear forms, and symmetric, skew-symmetric, and alternating bilinear forms).
positive definite means for every (Positive and negative definiteness, the inertia , rank , and signature of a real symmetric bilinear or quadratic form).
The addition formulas and the Pythagorean identity hold for all reals: , , and is even (The addition formulas for sine and cosine, Parity and the Pythagorean identity for sine and cosine).
For every real one has , and the Chebyshev polynomials of the first kind satisfy , and ( and for every , Chebyshev polynomials of the first and second kinds by their three-term recurrences).
Cosine is strictly decreasing on , is the smallest positive zero of cosine, and and are defined by their power series (Signs, monotonicity intervals, and ranges of sine and cosine, Pi as twice the smallest positive zero of cosine, Sine and cosine defined by their real power series).
Every has a unique with , and squaring is strictly increasing on the nonnegative reals (Square roots exist: a unique with ; the positives are , Squaring is monotone on the nonnegatives).
If is a subspace of a finite-dimensional inner product space , then , the orthogonal projection is the map with , and if then with precisely when (Real and complex inner-product spaces and their induced length, The orthogonal projection is the -component in , For a subspace of a finite-dimensional inner product space, ).
For all vectors , , with equality if and only if are linearly dependent (Cauchy–Schwarz: , with equality exactly for linearly dependent vectors).
The minor is the determinant of the matrix obtained by deleting row and column , the cofactor is , and determinant is the unique normalized alternating column-multilinear function of the matrix (Deleted-row-and-column minors, cofactors, the cofactor matrix and the adjugate over a commutative ring, The determinant is the unique normalized alternating multilinear function on the columns, Laplace expansion computes the determinant along every row and every column over a commutative ring).
The form a basis of , so vectors supported on disjoint subsets of are linearly independent unless one of them is , and subspaces are the published notions, and is a finite cardinality (Basis of a vector space: a linearly independent spanning subset; and ordered basis: an injective finite list whose image is a basis, Linear combination of a finite list, and the span as the smallest linear subspace containing , Linear subspace of a vector space, The cardinality of a finite set).
The induction principle on (The principle of mathematical induction, The natural numbers (von Neumann)).
Proof
(Trigonometric values and comparisons.) From [F4] one gets the double-angle formula and the triple-angle formula for every real . Cosine is strictly decreasing on with and is positive on [F5, F7]: (i) because and [F8]; (ii) writing , the triple-angle formula at gives , and because forces ; (iii) because and [F8]; (iv) since is strictly increasing in : for one has , for one has , and for one has . Also : with , iterating the recurrence of [F6] gives , , and , so [F5, F6], that is by expansion; since gives [F5, F7], one has and , i.e. ; by uniqueness of the nonnegative square root , and excludes the negative alternative, which is because by and [F8]; hence and satisfies because [F8].
(Witness principle and form comparison.) By [F3], positive definite means for every nonzero ; a nonzero with therefore contradicts positive definiteness, which is the first assertion of (1). For the comparison, let have all and let be a labelled graph on whose labels are at most the corresponding labels of ; for distinct the comparison of labels gives (with value exactly for the non-edges and throughout), so each term being ; hence , and if then with , excluding positive definiteness.
(Path determinant recursion.) Let be a path with vertices and , and let be the cosine matrix . Its leading submatrix has diagonal entries , sub- and super-diagonal entries , and all other entries . Put , with and . Expanding along its last row for : the last entry contributes , and writing , the other entry contributes , where is obtained by deleting row and column . Its last column has the sole nonzero entry in its last row, so expansion gives ; this contribution is therefore , giving
(Chain inequality.) Let be a path, all of whose labels are except one edge labelled , with ; by [F3] and [F9] the form makes a finite-dimensional inner product space. Put and , so that is supported on , on , and the coefficients of the two endpoints of the large edge are and . All internal edges of the two chains have label , and by step 1.1, so the diagonal terms and the two symmetric terms for each internal edge give with the same computation for , while every mixed pair contributes except , giving . Since are nonzero and supported on disjoint subsets of the basis , they are linearly independent [F12], so [F10] is strict, , that is , which gives .
(No cycles.) Let be distinct vertices whose consecutive pairs are edges, and put , a vector with all coordinates in the basis . Every consecutive pair contributes because on edges [1.1], and every other pair contributes (the value for non-edges, for the remaining edges), so ; by the witness principle [1.2] this contradicts positive definiteness of . Hence contains no cycle. Connectedness gives a path between any two vertices, and two different simple paths would give a cycle between their first divergence and subsequent reunion; thus that path is unique. Removing a vertex separates its neighbours into distinct components, since a path between two neighbours avoiding the removed vertex would create a cycle. Finally, a finite connected acyclic graph of maximum degree at most is a path: a longest simple path has no extension at either end, and an additional vertex would have a path to it meeting an internal vertex (creating degree at least ) or an endpoint (extending it). These elementary consequences will be used below.
(Path with all labels .) For the path with every label the recursion of [1.3] reads ; by induction on [F13] one has for all , since , and . In particular , and multiplying the matrix by scales its determinant by , so .
(Integer consequences of the chain inequality.) Write with , where means [F2]. If , then and , because . If , including , then by [1.1], so and the inequality forces ; then it reads , i.e. because , and hence (for it reads , false, so no infinite label occurs). If , then and by [1.1]; for the inequality fails because for , so , and then reads , so . If , then by [1.1] and the inequality reads : for this holds for every , for it reads , so , and for it fails since .
(Two large labels are impossible.) Suppose two distinct edges of carry labels . Since [2.2] shows that is acyclic, the two edges are joined by a unique simple path; write its vertices as so that and are the two large edges, . Put (all coordinates ). The diagonal contribution is . The two large edges contribute each, by [1.1] and ; each of the remaining path edges contributes ; and all other pairs contribute . Hence , so by the witness principle [1.2] is not positive definite, a contradiction. Therefore at most one edge of has label .
(The neighbour inequality.) Fix and let be neighbours of . If were joined by an edge, then would be three distinct vertices whose consecutive pairs are edges, i.e. a cycle, which [2.2] excludes; so for distinct neighbours. By [F2] each , and the restriction of to the subspace is an inner product, because is positive definite on [F3] and restricts to . Let with , be the orthogonal decomposition of [F9]; then and . Since together with the , , consists of distinct vectors of the basis , the vector is not in [F12], so and
(Three-arm inequality.) Suppose has degree and every edge of has label , and the three components of are paths with vertices. Weight each arm from its far end toward : if the arm of with vertices is with adjacent to , put , and define for the other two arms cyclically. Then by the same computation as in [2.1], since only the pair contributes, and are pairwise orthogonal because their supports lie in different components of , so no edge joins two of them [2.2]. Moreover , as involve only basis vectors different from ; hence the orthogonal decomposition of with respect to that subspace is strict and [F9] gives i.e. .
(Local consequences.) Fix and use [3.3] together with for every neighbour and . A neighbour with would have and hence a sum , so no label is ; four neighbours would give a sum , so ; if there are three neighbours and one of their edges has label , that edge contributes at least while the other two each contribute at least by step 1.1, giving a sum , a contradiction; hence all three edges have label ; and two neighbours with labels and satisfy : if then and (cosine increases with [1.1]), a contradiction, so and then ; since and cosine increases with , would give , so . In particular gives and gives , both excluded.
(Integer consequences of the three-arm inequality.) Let with . If , then and the sum is at most , so . Then : if this holds for every ; if it reads , i.e. , so ; and if the sum is at most , a contradiction. Hence, up to permutation, for some , or .
(At most one vertex of degree .) Suppose are two vertices of degree . By [4.1] both have degree exactly . By [2.2] the diagram is acyclic, so and are joined by a unique simple path with , the remaining neighbours of and of lie outside this path, and all of are pairwise distinct (two of them equal would create a second - path, or a triangle when ). Put , a nonzero vector with all coordinates : its diagonal contribution is , the path edges contribute each, the four pendant edges contribute each, and all other pairs contribute ; hence , contradicting positive definiteness by the witness principle [1.2]. So at most one vertex of degree exists, and by [4.1] at most one of degree .
(A degree- vertex excludes every large label.) Suppose has degree and some edge has label . Removing from the acyclic graph of [2.2] leaves three components, and the large edge lies in one of them: write in order along a path, so that is the large edge (possibly with a neighbour of ), and let be the neighbours of in the other two components. Then and are pairwise distinct except for the described edges, and is nonzero with all coordinates ; its diagonal contribution is with , the path edges contribute each, the large edge contributes , the two edges at toward contribute each, and all other pairs contribute . Hence because gives and by [1.1]; by the witness principle [1.2] this contradicts positive definiteness. So if a large label exists, no vertex of degree exists; combined with [4.1] every degree is and the acyclic connected is a path.
(Conclusion.) Let be connected and positive definite. By [2.2] it is acyclic, hence a tree. If has no vertex of degree , then by [4.1] all degrees are , so is a path: with all labels it is () [2.3]; otherwise by [3.2] exactly one edge has label and [5.2] applies, and splitting the path at that edge into subpaths of vertices, the inequality of [2.1] and its case analysis [3.1] give with , giving the single edge ; with or , giving and ; with and , giving the paths with labels (type , the label- edge at an end); and with , giving the paths with labels (type ) and (type ); and with , giving the path with labels (type ). If has a vertex of degree , it is unique by [5.1], its edges have label and no edge has label by [5.2] and [4.1], and the three arms have vertices satisfying the inequality of [3.4]; by [4.2] they are with , giving the star with arms , i.e. (), or , , , giving , , . Thus every connected positive definite diagram is one of (), (), (), , , , , () with the labels displayed in (7), which is the asserted list.
Depends on
- Coxeter diagrams: edges, labels, components and finite type
- The real Coxeter form, its radical, reflections, and form-preserving maps
- Positive and negative definiteness, the inertia $(p,q,r)$, rank $p+q$, and signature $p-q$ of a real symmetric bilinear or quadratic form
- Sylvester's criterion: a real symmetric $n\times n$ matrix with $n\geq1$ is positive definite if and only if all leading principal minors are positive
- Deleted-row-and-column minors, cofactors, the cofactor matrix and the adjugate over a commutative ring
- The determinant is the unique normalized alternating multilinear function on the columns
- Real and complex inner-product spaces and their induced length
- The orthogonal projection $P_Wv$ is the $W$-component in $V=W\oplus W^\perp$
- For a subspace $W$ of a finite-dimensional inner product space, $V=W\oplus W^\perp$
- Cauchy–Schwarz: $|\langle u,v\rangle|\leq\lVert u\rVert\lVert v\rVert$, with equality exactly for linearly dependent vectors
- Bilinear forms, and symmetric, skew-symmetric, and alternating bilinear forms
- Linear combination of a finite list, and the span $\operatorname{span}(S)$ as the smallest linear subspace containing $S$
- Basis of a vector space: a linearly independent spanning subset; and ordered basis: an injective finite list whose image is a basis
- Linear subspace of a vector space
- Sine and cosine defined by their real power series
- Pi as twice the smallest positive zero of cosine
- Quarter-turn values and shifts by pi/2 and pi
- $T_n(\cos\theta)=\cos(n\theta)$ and $U_n(\cos\theta)\sin\theta=\sin((n+1)\theta)$ for every $n\in\mathbb N$
- Chebyshev polynomials of the first and second kinds by their three-term recurrences
- The addition formulas for sine and cosine
- Parity and the Pythagorean identity for sine and cosine
- Signs, monotonicity intervals, and ranges of sine and cosine
- Square roots exist: a unique $\sqrt{a} \ge 0$ with $(\sqrt{a})^2 = a$; the positives are $\{x^2 : x \neq 0\}$
- Squaring is monotone on the nonnegatives
- The natural numbers $\mathbb{N}$ (von Neumann)
- The principle of mathematical induction
- The cardinality $\lvert A\rvert$ of a finite set
- Laplace expansion computes the determinant along every row and every column over a commutative ring
Used by
- The bipartite Coxeter element, its ordered prefix roots, and the conditional vector map mu(a) = -2(c-1)⁻¹a Definition
- A cycle and an overlong arm: explicit non-positive witnesses Example
- Bₙ and Cₙ define the same Coxeter diagram and the same Coxeter group Example
- Gram determinants and principal minors of the non-crystallographic types H₃ and H₄ Example
- Path determinants dₖ=dₖ₋₁-cos²(π/m)dₖ₋₂ and the three-arm inequality Example
- Cartan-number products, allowed edge labels, tree scalings and reflection stability Lemma
- Coxeter elements of tree type are conjugate by source and sink firings Lemma
- Enumeration of the connected positive semidefinite corank-one diagrams Lemma
- Classification of finite Coxeter systems, including the H and dihedral families Theorem
- Crystallographic finite type: the Weyl types, reduced realizations and lattice stability Theorem
Dependency tree · two levels
102 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.
Sources
- Jean Michel, Lectures on Coxeter groups (Beijing lecture notes, April-May 2014) (standard reference, not scraped)
- Michael W. Davis, The Geometry and Topology of Coxeter Groups (first-edition author manuscript, 2007-2008) (standard reference, not scraped)