{
  "id": "minimal-polynomial",
  "title": "Minimal Polynomial",
  "lang": "en",
  "source_lang": "en",
  "summary": "A minimal polynomial is the unique monic polynomial of least degree that annihilates an algebraic element, square matrix, or linear operator over a specified field.",
  "category": "mathematics",
  "subcategory": "algebra",
  "type": "concept",
  "aliases": [
    "Minimum polynomial"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/minimal-polynomial/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/minimal-polynomial/",
      "json": "https://aiwiki.page/zh/minimal-polynomial.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A minimal polynomial is a polynomial that expresses the lowest-degree polynomial relation satisfied by a given mathematical object over a specified field. For an algebraic element \\(\\alpha\\), it is the monic polynomial of least positive degree with \\(\\alpha\\) as a root. For a square matrix or a linear operator \\(T\\) on a nonzero finite-dimensional space, it is the monic polynomial of least degree that becomes zero when evaluated at \\(T\\). These definitions share a common algebraic foundation, but an important distinction separates them: the minimal polynomial of an algebraic element over a field is always irreducible, whereas the minimal polynomial of a matrix need not be."
    },
    {
      "heading": "General definition",
      "text": "Let \\(F\\) be a field and \\(a\\) an element of a nonzero unital associative \\(F\\)-algebra, with scalars acting centrally. For\n\n\\[p(x)=c_0+c_1x+\\cdots+c_dx^d\\in F[x],\\]\n\nevaluation at \\(a\\) means\n\n\\[p(a)=c_0\\,1+c_1a+\\cdots+c_da^d.\\]\n\nIf some nonzero polynomial satisfies \\(p(a)=0\\), then \\(a\\) is called algebraic over \\(F\\). Its minimal polynomial \\(m_a(x)\\) is the monic polynomial of smallest degree satisfying\n\n\\[m_a(a)=0.\\]\n\nHere monic means that the leading coefficient is \\(1\\).\n\nThe defining property implies the stronger divisibility statement\n\n\\[p(a)=0 \\quad\\Longleftrightarrow\\quad m_a(x)\\mid p(x).\\]\n\nIndeed, polynomial division gives \\(p=qm_a+r\\), where \\(\\deg r<\\deg m_a\\). If \\(p(a)=0\\), then \\(r(a)=0\\); minimality forces \\(r=0\\). This also proves uniqueness.\n\nEquivalently, the evaluation map\n\n\\[\\operatorname{ev}_a:F[x]\\longrightarrow F[a], \\qquad p\\longmapsto p(a),\\]\n\nis a ring homomorphism whose kernel is the ideal\n\n\\[\\ker(\\operatorname{ev}_a)=(m_a).\\]\n\nThus the minimal polynomial encodes every polynomial relation satisfied by \\(a\\), not merely one relation of smallest degree."
    },
    {
      "heading": "Algebraic elements and field extensions",
      "text": "Suppose \\(E/F\\) is a field extension and \\(\\alpha\\in E\\) is algebraic over \\(F\\). Its minimal polynomial is often denoted\n\n\\[m_{\\alpha,F}(x) \\quad\\text{or}\\quad \\operatorname{min}_F(\\alpha).\\]\n\nIrreducibility\n\nThe polynomial \\(m_{\\alpha,F}\\) is an irreducible polynomial over \\(F\\). If it factored as\n\n\\[m_{\\alpha,F}=gh\\]\n\nwith both factors of positive degree, then\n\n\\[0=g(\\alpha)h(\\alpha).\\]\n\nBecause \\(E\\) is a field, one factor would vanish at \\(\\alpha\\), contradicting minimality.\n\nConversely, any monic irreducible polynomial over \\(F\\) having \\(\\alpha\\) as a root is its minimal polynomial.\n\nDegree and a basis for the generated field\n\nThe evaluation map gives an isomorphism\n\n\\[F[x]/(m_{\\alpha,F})\\cong F[\\alpha].\\]\n\nIrreducibility makes this quotient a field, so \\(F[\\alpha]=F(\\alpha)\\), the smallest subfield of \\(E\\) containing \\(F\\) and \\(\\alpha\\).\n\nIf \\(d=\\deg m_{\\alpha,F}\\), then\n\n\\[[F(\\alpha):F]=d,\\]\n\nand\n\n\\[1,\\alpha,\\ldots,\\alpha^{d-1}\\]\n\nform a basis over \\(F\\). Every element of \\(F(\\alpha)\\) therefore has a unique expression as a polynomial in \\(\\alpha\\) of degree less than \\(d\\).\n\nExamples and dependence on the base field\n\nOver the rational numbers,\n\n\\[m_{\\sqrt2,\\mathbb Q}(x)=x^2-2.\\]\n\nThis polynomial vanishes at \\(\\sqrt2\\) and is irreducible over \\(\\mathbb Q\\).\n\nThe base field is essential. Since \\(\\sqrt2\\in\\mathbb R\\),\n\n\\[m_{\\sqrt2,\\mathbb R}(x)=x-\\sqrt2.\\]\n\nLikewise, the imaginary unit has minimal polynomial \\(x^2+1\\) over \\(\\mathbb R\\), but \\(x-i\\) over \\(\\mathbb C\\).\n\nA transcendental number over \\(F\\) satisfies no nonzero polynomial over \\(F\\), so it has no minimal polynomial over that field.\n\nConjugate roots and separability\n\nThe roots of \\(m_{\\alpha,F}\\) in a field where it splits are the algebraic conjugates of \\(\\alpha\\) over \\(F\\). They are central to Galois theory.\n\nIn characteristic zero, an irreducible polynomial has no repeated roots. In positive characteristic, an irreducible polynomial may have repeated roots in a splitting field. Thus irreducibility does not by itself imply separability over every field."
    },
    {
      "heading": "Matrices and linear operators",
      "text": "Let \\(V\\) be a nonzero finite-dimensional vector space over \\(F\\), and let \\(T:V\\to V\\) be a linear operator. For\n\n\\[p(x)=c_0+c_1x+\\cdots+c_dx^d,\\]\n\ndefine\n\n\\[p(T)=c_0I+c_1T+\\cdots+c_dT^d,\\]\n\nwhere \\(I\\) is the identity operator.\n\nThe minimal polynomial \\(m_T\\) is the unique monic polynomial of least degree satisfying \\(m_T(T)=0\\). It exists because sufficiently many powers of \\(T\\) are linearly dependent in the finite-dimensional space of operators on \\(V\\).\n\nChoosing a basis gives a matrix \\(A\\) representing \\(T\\), and \\(m_A=m_T\\). In particular, the minimal polynomial is invariant under matrix similarity:\n\n\\[B=P^{-1}AP \\quad\\Longrightarrow\\quad m_B=m_A.\\]\n\nUnlike the field-element case, an operator acts in an algebra that may have zero divisors. Its minimal polynomial can therefore factor into several nonconstant polynomials.\n\nRelation to the characteristic polynomial\n\nFor an \\(n\\times n\\) matrix \\(A\\), the characteristic polynomial is\n\n\\[\\chi_A(x)=\\det(xI-A).\\]\n\nThe Cayley–Hamilton theorem states that\n\n\\[\\chi_A(A)=0.\\]\n\nConsequently,\n\n\\[m_A\\mid\\chi_A, \\qquad 1\\leq\\deg m_A\\leq n.\\]\n\nThe two polynomials contain different information. The characteristic polynomial records eigenvalues with their algebraic multiplicities. The minimal polynomial records the polynomial relations needed to annihilate the operator and, over a splitting field, the largest Jordan block associated with each eigenvalue.\n\nThey have the same distinct roots in a splitting field, but root multiplicities can differ. For example,\n\n\\[A=\\begin{pmatrix}2&0\\\\0&2\\end{pmatrix}\\]\n\nhas\n\n\\[\\chi_A(x)=(x-2)^2, \\qquad m_A(x)=x-2.\\]\n\nEigenvalues and Jordan blocks\n\nAn element \\(\\lambda\\) in a splitting field is an eigenvalue of \\(A\\) exactly when\n\n\\[m_A(\\lambda)=0.\\]\n\nIf\n\n\\[m_A(x)=\\prod_{\\lambda}(x-\\lambda)^{s_\\lambda},\\]\n\nthen \\(s_\\lambda\\) is the size of the largest block for \\(\\lambda\\) in the Jordan normal form of \\(A\\).\n\nFor example,\n\n\\[B=\\begin{pmatrix}2&1\\\\0&2\\end{pmatrix}\\]\n\nsatisfies\n\n\\[(B-2I)^2=0, \\qquad B-2I\\neq0,\\]\n\nso\n\n\\[m_B(x)=(x-2)^2.\\]\n\nAlthough \\(A\\) and \\(B\\) above have the same characteristic polynomial, their minimal polynomials distinguish their different operator structures.\n\nMore generally, if \\(N^r=0\\) but \\(N^{r-1}\\neq0\\), then \\(N\\) is nilpotent of index \\(r\\), and\n\n\\[m_N(x)=x^r.\\]\n\nDiagonalizability\n\nA matrix is diagonalizable over \\(F\\) if and only if its minimal polynomial is a product of distinct linear factors in \\(F[x]\\).\n\nBoth requirements matter:\n\n- Repeated factors indicate nontrivial Jordan blocks.\n- Factors that do not split over \\(F\\) prevent diagonalization over \\(F\\), even if no irreducible factor is repeated.\n\nFor example, the real matrix\n\n\\[R=\\begin{pmatrix}0&-1\\\\1&0\\end{pmatrix}\\]\n\nhas minimal polynomial \\(x^2+1\\). It is diagonalizable over \\(\\mathbb C\\), but not over \\(\\mathbb R\\).\n\nFor a matrix whose entries lie in \\(F\\), extending scalars to a larger field does not change its minimal polynomial as a polynomial; only its factorization may change. This differs from fixing an element of an extension field and enlarging the base field over which that element’s minimal polynomial is taken."
    },
    {
      "heading": "Invariant subspaces and canonical forms",
      "text": "If \\(W\\subseteq V\\) is invariant under \\(T\\), then\n\n\\[m_{T|_W}\\mid m_T.\\]\n\nIf\n\n\\[V=V_1\\oplus\\cdots\\oplus V_k\\]\n\nis a decomposition into invariant subspaces, then\n\n\\[m_T=\\operatorname{lcm}\\bigl(m_{T|_{V_1}},\\ldots,m_{T|_{V_k}}\\bigr).\\]\n\nA useful interpretation makes \\(V\\) an \\(F[x]\\)-module by defining \\(xv=T(v)\\). In rational canonical form, the operator has monic nonconstant invariant factors\n\n\\[f_1\\mid f_2\\mid\\cdots\\mid f_r.\\]\n\nThey satisfy\n\n\\[\\chi_T=\\prod_{j=1}^r f_j, \\qquad m_T=f_r.\\]\n\nThus the minimal polynomial is the largest invariant factor. It need not determine the complete similarity class: different operators can have the same minimal polynomial but different remaining invariant factors.\n\nAn operator is cyclic if some vector \\(v\\) has\n\n\\[v,Tv,\\ldots,T^{n-1}v\\]\n\nas a basis. This happens exactly when\n\n\\[m_T=\\chi_T.\\]\n\nThe two principal meanings of minimal polynomial meet naturally here. If \\(\\alpha\\) is algebraic over \\(F\\), multiplication by \\(\\alpha\\) defines an operator on \\(F(\\alpha)\\):\n\n\\[M_\\alpha(u)=\\alpha u.\\]\n\nIts minimal polynomial is \\(m_{\\alpha,F}\\). The vector \\(1\\) is cyclic, so its characteristic polynomial is also \\(m_{\\alpha,F}\\)."
    },
    {
      "heading": "Computation and applications",
      "text": "Computing matrix minimal polynomials\n\nA direct method seeks the first linear dependence among\n\n\\[I,A,A^2,\\ldots.\\]\n\nIf \\(d\\) is the first degree at which a dependence occurs, it can be normalized as\n\n\\[A^d+c_{d-1}A^{d-1}+\\cdots+c_0I=0.\\]\n\nThen\n\n\\[m_A(x)=x^d+c_{d-1}x^{d-1}+\\cdots+c_0.\\]\n\nFor a vector \\(v\\), the first dependence among \\(v,Av,A^2v,\\ldots\\) gives its vector minimal polynomial \\(m_{A,v}\\). This divides \\(m_A\\) and describes the action on the subspace generated by those vectors. Taking the least common multiple of the vector minimal polynomials for a basis recovers \\(m_A\\).\n\nOther methods use invariant factors or factorization of the characteristic polynomial. Exact minimal-polynomial computation is fundamentally algebraic: small perturbations of a matrix can change its minimal polynomial, so exact structure should not be inferred uncritically from floating-point calculations.\n\nReducing powers and computing inverses\n\nIf \\(\\deg m_A=d\\), polynomial division shows that every polynomial in \\(A\\) can be written uniquely as a linear combination of\n\n\\[I,A,\\ldots,A^{d-1}.\\]\n\nThis reduces high powers to lower ones and yields recurrence relations for sequences generated by matrix powers.\n\nThe matrix \\(A\\) is invertible exactly when \\(m_A(0)\\neq0\\). If\n\n\\[m_A(x)=x^d+c_{d-1}x^{d-1}+\\cdots+c_1x+c_0\\]\n\nwith \\(c_0\\neq0\\), then\n\n\\[A^{-1} =-\\frac{1}{c_0} \\left(A^{d-1}+c_{d-1}A^{d-2}+\\cdots+c_1I\\right).\\]\n\nMatrix functions and field arithmetic\n\nOver \\(\\mathbb C\\), an analytic function of a matrix, such as the matrix exponential, can be represented by a polynomial in \\(A\\) of degree less than \\(\\deg m_A\\). The representing polynomial matches the function’s values at eigenvalues and the required derivatives determined by the multiplicities in \\(m_A\\).\n\nFor algebraic field elements, minimal polynomials support exact arithmetic by reducing expressions modulo the defining polynomial. Over a finite field with \\(q\\) elements, a monic irreducible polynomial \\(m\\) of degree \\(d\\) gives the field\n\n\\[\\mathbb F_q[x]/(m),\\]\n\nwhich has \\(q^d\\) elements. The residue class of \\(x\\) has minimal polynomial \\(m\\) over \\(\\mathbb F_q\\)."
    }
  ],
  "links": [
    "polynomial",
    "field-mathematics",
    "matrix",
    "algebra",
    "homomorphism",
    "ideal-ring-theory",
    "field-extension",
    "isomorphism",
    "basis-linear-algebra",
    "rational-number",
    "transcendental-number",
    "galois-theory",
    "vector-space",
    "linear-map",
    "matrix-similarity",
    "characteristic-polynomial",
    "cayley-hamilton-theorem",
    "eigenvalues-and-eigenvectors",
    "jordan-normal-form",
    "matrix-diagonalization",
    "module-mathematics",
    "recurrence-relation",
    "matrix-exponential",
    "finite-field"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}