{
  "id": "skew-symmetric-matrix",
  "title": "Skew-Symmetric Matrix",
  "lang": "en",
  "source_lang": "en",
  "summary": "A skew-symmetric matrix equals the negative of its transpose; over the real numbers, such matrices describe infinitesimal rotations and alternating bilinear forms.",
  "category": "mathematics",
  "subcategory": "algebra",
  "type": "concept",
  "aliases": [
    "Antisymmetric matrix"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/skew-symmetric-matrix/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/skew-symmetric-matrix/",
      "json": "https://aiwiki.page/zh/skew-symmetric-matrix.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A skew-symmetric matrix is a square matrix \\(A\\) whose transpose equals its negative:\n\n\\[A^{\\mathsf T}=-A.\\]\n\nEquivalently, its entries satisfy \\(a_{ij}=-a_{ji}\\). Over the real or complex numbers, every diagonal entry is therefore zero. Skew-symmetric matrices are fundamental in linear algebra, where they represent alternating bilinear forms and, over the real numbers, infinitesimal rotations. Their structure imposes strong restrictions on their rank, determinant, and eigenvalues."
    },
    {
      "heading": "Definition and examples",
      "text": "For a matrix over a field of characteristic other than \\(2\\), the diagonal condition follows from\n\n\\[a_{ii}=-a_{ii} \\quad\\Longrightarrow\\quad 2a_{ii}=0.\\]\n\nA general \\(3\\times3\\) skew-symmetric matrix has the form\n\n\\[A= \\begin{pmatrix} 0&a&b\\\\ -a&0&c\\\\ -b&-c&0 \\end{pmatrix}.\\]\n\nThe entries above the diagonal can be chosen freely; those below it are then determined.\n\nThe simplest nonzero example is\n\n\\[J= \\begin{pmatrix} 0&-1\\\\ 1&0 \\end{pmatrix}.\\]\n\nAs a linear transformation of the plane, \\(J\\) rotates vectors counterclockwise through a right angle, and \\(J^2=-I\\).\n\nThe only matrix that is both skew-symmetric and symmetric is the zero matrix, provided the underlying field has characteristic other than \\(2\\)."
    },
    {
      "heading": "Vector-space structure and decomposition",
      "text": "The \\(n\\times n\\) skew-symmetric matrices form a vector space: sums and scalar multiples remain skew-symmetric. Its dimension is\n\n\\[\\frac{n(n-1)}2,\\]\n\nthe number of entries strictly above the diagonal. A standard basis consists of\n\n\\[E_{ij}-E_{ji},\\qquad 1\\leq i<j\\leq n,\\]\n\nwhere \\(E_{ij}\\) has a single \\(1\\) in position \\((i,j)\\) and zeros elsewhere.\n\nEvery square matrix \\(M\\) over a field of characteristic other than \\(2\\) has the unique decomposition\n\n\\[M=S+K, \\qquad S=\\frac{M+M^{\\mathsf T}}2, \\qquad K=\\frac{M-M^{\\mathsf T}}2,\\]\n\nwith \\(S\\) symmetric and \\(K\\) skew-symmetric.\n\nFor real matrices, these two subspaces are orthogonal under the Frobenius inner product\n\n\\[\\langle X,Y\\rangle_F=\\operatorname{tr}(X^{\\mathsf T}Y).\\]\n\nConsequently, \\((M-M^{\\mathsf T})/2\\) is the orthogonal projection of \\(M\\) onto the skew-symmetric subspace, and is the closest skew-symmetric matrix to \\(M\\) in the Frobenius norm.\n\nSkew-symmetric matrices are not generally closed under matrix multiplication. In fact, for skew-symmetric \\(A\\) and \\(B\\),\n\n\\[(AB)^{\\mathsf T}=BA.\\]\n\nThus \\(AB\\) is symmetric if \\(A\\) and \\(B\\) commute. Their commutator,\n\n\\[[A,B]=AB-BA,\\]\n\nis always skew-symmetric."
    },
    {
      "heading": "Bilinear forms and quadratic expressions",
      "text": "A matrix \\(A\\) determines a bilinear form\n\n\\[b(x,y)=x^{\\mathsf T}Ay.\\]\n\nIf \\(A\\) is skew-symmetric, then\n\n\\[b(y,x)=-b(x,y).\\]\n\nOver a field of characteristic other than \\(2\\), this is equivalent to the form being alternating, meaning\n\n\\[b(x,x)=0\\]\n\nfor every vector \\(x\\).\n\nIn particular,\n\n\\[x^{\\mathsf T}Ax=0\\]\n\nfor every real vector \\(x\\). Geometrically, \\(Ax\\) is perpendicular to \\(x\\) in the standard Euclidean inner product.\n\nThis identity also explains why a quadratic form depends only on the symmetric part of its coefficient matrix:\n\n\\[x^{\\mathsf T}Mx = x^{\\mathsf T}\\left(\\frac{M+M^{\\mathsf T}}2\\right)x.\\]\n\nThe skew-symmetric part contributes nothing.\n\nUnder a change of coordinates \\(x=Pu\\), the matrix of the bilinear form becomes\n\n\\[P^{\\mathsf T}AP,\\]\n\nwhich remains skew-symmetric. This transformation is a congruence, not generally a similarity transformation. For orthogonal \\(P\\), the two coincide because \\(P^{\\mathsf T}=P^{-1}\\)."
    },
    {
      "heading": "Rank, determinant, and the Pfaffian",
      "text": "If \\(A\\) has odd order \\(n\\), then\n\n\\[\\det A =\\det A^{\\mathsf T} =\\det(-A) =(-1)^n\\det A =-\\det A.\\]\n\nOver a field of characteristic other than \\(2\\), this implies\n\n\\[\\det A=0.\\]\n\nEvery odd-order skew-symmetric matrix is therefore singular.\n\nMore generally, the rank of a skew-symmetric matrix over such a field is always even. Under an invertible congruence transformation, it can be reduced to a block diagonal matrix consisting of blocks\n\n\\[\\begin{pmatrix} 0&1\\\\ -1&0 \\end{pmatrix}\\]\n\nand a zero block. Each nonzero block contributes rank \\(2\\).\n\nFor an even-order skew-symmetric matrix, its determinant is the square of a polynomial in its entries called the Pfaffian:\n\n\\[\\det A=\\operatorname{pf}(A)^2.\\]\n\nFor example, if\n\n\\[A= \\begin{pmatrix} 0&a&b&c\\\\ -a&0&d&e\\\\ -b&-d&0&f\\\\ -c&-e&-f&0 \\end{pmatrix},\\]\n\nthen\n\n\\[\\operatorname{pf}(A)=af-be+cd, \\qquad \\det A=(af-be+cd)^2.\\]\n\nIn particular, every real skew-symmetric matrix has nonnegative determinant, including zero in the singular case."
    },
    {
      "heading": "Eigenvalues and real canonical form",
      "text": "A real skew-symmetric matrix satisfies\n\n\\[A^*=-A,\\]\n\nwhere \\(A^*\\) denotes the conjugate transpose. It is therefore a special case of a skew-Hermitian matrix.\n\nEvery eigenvalue of a real skew-symmetric matrix is purely imaginary or zero. Nonzero eigenvalues occur in conjugate pairs\n\n\\[+i\\sigma,\\qquad -i\\sigma,\\]\n\nwith \\(\\sigma>0\\).\n\nSuch a matrix is normal, so the complex spectral theorem gives a unitary diagonalization. Over the real numbers, the corresponding result is an orthogonal block decomposition: there exists an orthogonal matrix \\(Q\\) such that\n\n\\[Q^{\\mathsf T}AQ = \\operatorname{diag}\\left( J(\\sigma_1),\\ldots,J(\\sigma_r),0_{n-2r} \\right),\\]\n\nwhere\n\n\\[J(\\sigma)= \\begin{pmatrix} 0&-\\sigma\\\\ \\sigma&0 \\end{pmatrix}, \\qquad \\sigma_j>0.\\]\n\nEach block acts in an invariant two-dimensional plane, while the zero block acts on the null space. This decomposition makes the even-rank property explicit. It also shows that each positive singular value occurs twice.\n\nFurthermore,\n\n\\[A^{\\mathsf T}A=-A^2.\\]\n\nThus \\(-A^2\\) is symmetric positive semidefinite, since\n\n\\[x^{\\mathsf T}(-A^2)x=\\|Ax\\|^2\\geq0.\\]"
    },
    {
      "heading": "Rotations and matrix exponentials",
      "text": "For a real skew-symmetric matrix \\(A\\), its matrix exponential\n\n\\[e^{tA}=\\sum_{k=0}^{\\infty}\\frac{t^kA^k}{k!}\\]\n\nis orthogonal for every real \\(t\\):\n\n\\[(e^{tA})^{\\mathsf T}e^{tA} =e^{-tA}e^{tA} =I.\\]\n\nIts determinant is \\(1\\), so it is a rotation in the higher-dimensional sense.\n\nFor a single canonical block,\n\n\\[e^{tJ(\\sigma)} = \\begin{pmatrix} \\cos(\\sigma t)&-\\sin(\\sigma t)\\\\ \\sin(\\sigma t)&\\cos(\\sigma t) \\end{pmatrix}.\\]\n\nA general real skew-symmetric matrix therefore generates simultaneous rotations in mutually orthogonal planes, leaving its null space fixed.\n\nConversely, if \\(Q(t)\\) is a differentiable curve of orthogonal matrices with \\(Q(0)=I\\), differentiating\n\n\\[Q(t)^{\\mathsf T}Q(t)=I\\]\n\nat \\(t=0\\) shows that \\(Q'(0)\\) is skew-symmetric.\n\nThe real skew-symmetric matrices, equipped with the commutator, form the Lie algebra \\(\\mathfrak{so}(n)\\) of the rotation Lie group \\(SO(n)\\). This is the precise sense in which they describe infinitesimal rotations.\n\nA related construction is the Cayley transform\n\n\\[Q=(I-A)^{-1}(I+A).\\]\n\nFor real skew-symmetric \\(A\\), \\(I-A\\) is invertible and \\(Q\\) is orthogonal with determinant \\(1\\). Unlike the exponential map, this parametrization cannot produce orthogonal matrices having eigenvalue \\(-1\\)."
    },
    {
      "heading": "Three-dimensional cross products",
      "text": "For \\(a=(a_1,a_2,a_3)^{\\mathsf T}\\), define\n\n\\[[a]_\\times= \\begin{pmatrix} 0&-a_3&a_2\\\\ a_3&0&-a_1\\\\ -a_2&a_1&0 \\end{pmatrix}.\\]\n\nThen\n\n\\[[a]_\\times x=a\\times x,\\]\n\nwhere \\(\\times\\) is the cross product.\n\nEvery real \\(3\\times3\\) skew-symmetric matrix has this form for a unique vector \\(a\\). When \\(a\\neq0\\), its null space is the line spanned by \\(a\\), and its eigenvalues are\n\n\\[0,\\quad i\\|a\\|,\\quad -i\\|a\\|.\\]\n\nThe exponential \\(e^{t[a]_\\times}\\) is rotation about the axis \\(a\\) through angle \\(t\\|a\\|\\), with orientation determined by the right-hand rule.\n\nThis correspondence is widely used in rigid-body kinematics, robotics, and classical mechanics to express angular velocity as a linear operation on position vectors."
    },
    {
      "heading": "Complex matrices and characteristic two",
      "text": "Two qualifications are essential when extending the real theory.\n\nComplex entries. A complex skew-symmetric matrix satisfies \\(A^{\\mathsf T}=-A\\), whereas a skew-Hermitian matrix satisfies \\(A^*=-A\\). These are different conditions. The purely imaginary eigenvalue property belongs to skew-Hermitian matrices and need not hold for complex skew-symmetric matrices. For example,\n\n\\[A= \\begin{pmatrix} 0&i\\\\ -i&0 \\end{pmatrix}\\]\n\nis complex skew-symmetric but has eigenvalues \\(1\\) and \\(-1\\).\n\nCharacteristic two. In a field of characteristic \\(2\\), \\(-A=A\\), so the equation \\(A^{\\mathsf T}=-A\\) merely says that \\(A\\) is symmetric. It does not force the diagonal to vanish. To represent an alternating bilinear form in this setting, zero diagonal must be required separately. Alternating matrices still have even rank, but matrices satisfying the transpose equation alone need not."
    },
    {
      "heading": "Applications to differential equations",
      "text": "Consider the linear ordinary differential equation\n\n\\[\\dot{x}=Ax\\]\n\nwith real skew-symmetric \\(A\\). Its solution is \\(x(t)=e^{tA}x(0)\\), and\n\n\\[\\frac{d}{dt}\\|x(t)\\|^2 =x(t)^{\\mathsf T}(A^{\\mathsf T}+A)x(t) =0.\\]\n\nThe motion preserves Euclidean length rather than producing growth or decay. The same norm-preservation calculation holds for a time-dependent real skew-symmetric coefficient matrix \\(A(t)\\).\n\nThis structure is important in numerical analysis because a time-stepping method may fail to preserve the geometry of the exact flow. For example, forward Euler gives\n\n\\[x_{k+1}=(I+hA)x_k,\\]\n\nand hence\n\n\\[\\|x_{k+1}\\|^2 = \\|x_k\\|^2+h^2\\|Ax_k\\|^2,\\]\n\nso it generally increases the norm. The implicit midpoint method instead uses the update matrix\n\n\\[\\left(I-\\frac h2A\\right)^{-1} \\left(I+\\frac h2A\\right),\\]\n\na Cayley transform that is orthogonal in exact arithmetic."
    }
  ],
  "links": [
    "matrix",
    "matrix-transpose",
    "linear-algebra",
    "field-mathematics",
    "symmetric-matrix",
    "vector-space",
    "inner-product",
    "orthogonal-projection",
    "quadratic-form",
    "matrix-similarity",
    "matrix-rank",
    "determinant",
    "conjugate-transpose",
    "eigenvalues-and-eigenvectors",
    "spectral-theorem",
    "orthogonal-matrix",
    "null-space",
    "matrix-exponential",
    "lie-algebra",
    "lie-group",
    "cross-product",
    "classical-mechanics",
    "ordinary-differential-equation",
    "numerical-analysis"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}