Math 2.0

Geometry Proof

Proof of the Cartan–Dieudonné theorem by induction on the dimension

Builds on Inner productLinear isometryA linear isometry preserves the inner productReflectionOrthogonal projectionOrthogonal complementA space is the direct sum of a subspace and its orthogonal complementDimension of the orthogonal complementA subspace of full dimension is the whole space

Proof

We argue by induction on nn. A hyperplane of VV is a subspace of dimension n−1n - 1.

  1. Let VV be a real inner product space with inner product ⟨⋅,⋅⟩\ip{\cdot}{\cdot} and norm ∥⋅∥\norm{\cdot}. For u∈Vu \in V with u≠0u \ne 0, set Hu={x∈V:⟨x,u⟩=0}H_u = \set{x \in V : \ip{x}{u} = 0}. Over ℝ\R, the inner product is symmetric.
  2. HuH_u is the orthogonal complement of the line ℝu\R u. If dim⁡V=n\dim V = n, then dim⁡Hu=n−1\dim H_u = n - 1 by the dimension formula. So HuH_u is a hyperplane.
  3. Claim: the reflection in HuH_u is Ru(x)=x−2⟨x,u⟩⟨u,u⟩uR_u(x) = x - 2 \frac{\ip{x}{u}}{\ip{u}{u}} u. Set c=⟨x,u⟩/⟨u,u⟩c = \ip{x}{u} / \ip{u}{u} and p=x−cup = x - c u.
  4. Then ⟨p,u⟩=⟨x,u⟩−c⟨u,u⟩=0\ip{p}{u} = \ip{x}{u} - c \ip{u}{u} = 0, so p∈Hup \in H_u. For h∈Huh \in H_u, ⟨cu,h⟩=c⟨h,u⟩=0\ip{c u}{h} = c \ip{h}{u} = 0, so cu∈Hu⊥c u \in H_u^\perp.
  5. Since x=p+cux = p + c u and the decomposition V=Hu⊕Hu⊥V = H_u \oplus H_u^\perp is unique, the orthogonal projection is PHu(x)=pP_{H_u}(x) = p. So the reflection is 2p−x=x−2cu=Ru(x)2p - x = x - 2 c u = R_u(x).
  6. From step 3, Ru(u)=−uR_u(u) = -u and Ru(h)=hR_u(h) = h for h∈Huh \in H_u. Since x=p+cux = p + cu, we get Ru(Ru(x))=Ru(p−cu)=p+cu=xR_u(R_u(x)) = R_u(p - cu) = p + cu = x. So Ru∘Ru=id⁡R_u \circ R_u = \id.
  7. For n≥0n \ge 0, let P(n)P(n) say: every linear isometry of an nn-dimensional real inner product space is a composition of at most nn reflections in hyperplanes. Let VV have dimension nn and let T ⁣:V→VT \colon V \to V be a linear isometry.
  8. If T=id⁡T = \id, then TT is the composition of zero reflections. This covers n=0n = 0, where V={0}V = \set{0}. Now let n≥1n \ge 1, assume P(n−1)P(n - 1), and let T≠id⁡T \ne \id.
  9. Choose v∈Vv \in V with T(v)≠vT(v) \ne v, and set u=T(v)−v≠0u = T(v) - v \ne 0. By preservation of inner products, ⟨T(v),T(v)⟩=⟨v,v⟩\ip{T(v)}{T(v)} = \ip{v}{v}. Write a=⟨v,v⟩a = \ip{v}{v} and b=⟨T(v),v⟩b = \ip{T(v)}{v}.
  10. Then ⟨u,u⟩=a−2b+a=2(a−b)\ip{u}{u} = a - 2b + a = 2(a - b) and ⟨T(v),u⟩=a−b\ip{T(v)}{u} = a - b. Since u≠0u \ne 0, a−b≠0a - b \ne 0.
  11. By step 3, Ru(T(v))=T(v)−2a−b2(a−b)u=T(v)−u=vR_u(T(v)) = T(v) - 2 \frac{a - b}{2(a - b)} u = T(v) - u = v.
  12. Let S=Ru∘TS = R_u \circ T. A reflection keeps norms, since ∥p−cu∥2=∥p∥2+c2∥u∥2=∥p+cu∥2\norm{p - cu}^2 = \norm{p}^2 + c^2 \norm{u}^2 = \norm{p + cu}^2 when p⊥up \perp u. So SS is a linear isometry with S(v)=vS(v) = v.
  13. Since T(0)=0T(0) = 0, we have v≠0v \ne 0. Let E=HvE = H_v, a hyperplane by step 2. For x∈Ex \in E, preservation of inner products by SS gives ⟨S(x),v⟩=⟨S(x),S(v)⟩=⟨x,v⟩=0\ip{S(x)}{v} = \ip{S(x)}{S(v)} = \ip{x}{v} = 0. So S(E)⊆ES(E) \subseteq E.
  14. The restriction S∣E ⁣:E→ES|_E \colon E \to E is a linear isometry of the real inner product space EE, of dimension n−1n - 1.
  15. By P(n−1)P(n-1), S∣E=R1′∘⋯∘Rk′S|_E = R'_1 \circ \cdots \circ R'_k with k≤n−1k \le n - 1, where Rj′R'_j is the reflection of EE in a hyperplane GjG_j of EE.
  16. Fix jj. The orthogonal complement of GjG_j in EE has dimension (n−1)−(n−2)=1(n - 1) - (n - 2) = 1 by the dimension formula. Choose wj≠0w_j \ne 0 in it.
  17. Then Gj⊆Gj′={x∈E:⟨x,wj⟩=0}G_j \subseteq G'_j = \set{x \in E : \ip{x}{w_j} = 0}, and dim⁡Gj′=n−2\dim G'_j = n - 2 by step 2 in EE. A subspace of full dimension is the whole space, so Gj=Gj′G_j = G'_j.
  18. By steps 3 to 5 applied in EE, Rj′(x)=x−2⟨x,wj⟩⟨wj,wj⟩wjR'_j(x) = x - 2 \frac{\ip{x}{w_j}}{\ip{w_j}{w_j}} w_j for x∈Ex \in E.
  19. Let Rj=RwjR_j = R_{w_j} be the reflection of VV in HwjH_{w_j}, a hyperplane of VV. By step 3 and step 18, Rj(x)=Rj′(x)R_j(x) = R'_j(x) for x∈Ex \in E. Also Rj(v)=vR_j(v) = v, since ⟨v,wj⟩=0\ip{v}{w_j} = 0.
  20. Each RjR_j maps EE into EE. So R1∘⋯∘RkR_1 \circ \cdots \circ R_k agrees with SS on EE, and both fix vv.
  21. By the decomposition for the line ℝv\R v, each x∈Vx \in V is tv+et v + e with t∈ℝt \in \R and e∈Ee \in E, as E=(ℝv)⊥E = (\R v)^\perp. Two linear maps that agree on vv and on EE are equal.
  22. Hence S=R1∘⋯∘RkS = R_1 \circ \cdots \circ R_k. By step 6, T=Ru∘Ru∘T=Ru∘S=Ru∘R1∘⋯∘RkT = R_u \circ R_u \circ T = R_u \circ S = R_u \circ R_1 \circ \cdots \circ R_k.
  23. This is a composition of k+1≤nk + 1 \le n reflections in hyperplanes of VV. This proves P(n)P(n), and the theorem follows by induction.

Idea

If TT moves a vector vv, the reflection in the hyperplane orthogonal to T(v)−vT(v) - v swaps T(v)T(v) and vv, because they have the same length. Composing with it gives an isometry that fixes vv, and hence maps the hyperplane orthogonal to vv to itself. Induction on that hyperplane gives at most n−1n - 1 more reflections.

Reading path

To understand this page from the ground up, read these 95 pages in order. The foundations come first. The path takes the proofs Proof of the Steinitz exchange lemma by induction, Proof that a subspace of full dimension is the whole space, Proof of the existence of an orthonormal basis by the Gram–Schmidt process, Proof of Grassmann's formula by extending a basis of the intersection, Proof that the Gram–Schmidt vectors are orthogonal, Proof of the polarisation identity, Proof that an orthonormal family is linearly independent, Proof of the formula for the dimension of the orthogonal complement, Proof that every linearly independent set extends to a basis, by Zorn's lemma, Proof of the formula for coordinates in an orthonormal basis, Proof of Zorn's lemma, Proof that n independent or n spanning vectors in an n-dimensional space form a basis, Proof that a linear isometry preserves the inner product by polarisation, Proof that a space is the direct sum of a finite-dimensional subspace and its orthogonal complement, Proof that a subspace of a finite-dimensional space has smaller or equal dimension, Proof that an intersection of subspaces is a subspace.

  1. Complex numbers object
  2. Element definition
  3. Inner product concept
  4. Inner product space concept
  5. Norm induced by an inner product concept
  6. Proof of Zorn's lemma proof
  7. Proposition concept
  8. Vector space definition
  9. Real and imaginary parts definition
  10. Set definition
  11. Disjunction definition
  12. Implication definition
  13. Predicate definition
  14. Inner product definition
  15. Linear combination definition
  16. Linear map definition
  17. Linear subspace definition
  18. Norm definition
  19. Complex conjugate definition
  20. Family of sets definition
  21. Ordered pair definition
  22. Union definition
  23. Existential quantifier definition
  24. Universal quantifier definition
  25. Inner product space definition
  26. Orthogonal vectors definition
  27. Linear independence definition
  28. Endomorphism definition
  29. Linear span definition
  30. Proof that an intersection of subspaces is a subspace proof
  31. Sum of subspaces definition
  32. Normed vector space definition
  33. Cartesian product definition
  34. Proof of the polarisation identity proof
  35. Subset definition
  36. Union of a family of sets definition
  37. Uniqueness quantifier definition
  38. Orthonormal family definition
  39. Norm induced by an inner product definition
  40. Orthogonal complement definition
  41. Intersection of subspaces statement
  42. Linear isometry definition
  43. Spanning set definition
  44. Direct sum of subspaces definition
  45. Binary relation definition
  46. Polarisation identity statement
  47. Proof that an orthonormal family is linearly independent proof
  48. Gram–Schmidt process concept
  49. Basis definition
  50. Proof of the Steinitz exchange lemma by induction proof
  51. Finite-dimensional vector space definition
  52. Projection definition
  53. Antisymmetric relation definition
  54. Reflexive relation definition
  55. Transitive relation definition
  56. Function definition
  57. Dimension theorem for vector spaces statement
  58. Orthonormal basis definition
  59. An orthonormal family is linearly independent statement
  60. Dimension definition
  61. Proof that a linear isometry preserves the inner product by polarisation proof
  62. Proof that the Gram–Schmidt vectors are orthogonal proof
  63. Steinitz exchange lemma statement
  64. Preorder definition
  65. Choice function definition
  66. Proof of the formula for coordinates in an orthonormal basis proof
  67. A linear isometry preserves the inner product statement
  68. Gram–Schmidt vectors are orthogonal statement
  69. Proof that a subspace of a finite-dimensional space has smaller or equal dimension proof
  70. Proof that n independent or n spanning vectors in an n-dimensional space form a basis proof
  71. Partial order definition
  72. Axiom of choice statement
  73. Coordinates in an orthonormal basis statement
  74. Linearly independent set of the right size is a basis statement
  75. Dimension of a subspace statement
  76. Maximal element definition
  77. Total order definition
  78. Upper bound definition
  79. Proof that a subspace of full dimension is the whole space proof
  80. Chain definition
  81. A subspace of full dimension is the whole space statement
  82. Zorn's lemma statement
  83. Proof that every linearly independent set extends to a basis, by Zorn's lemma proof
  84. Every linearly independent set extends to a basis statement
  85. Proof of the existence of an orthonormal basis by the Gram–Schmidt process proof
  86. Proof of Grassmann's formula by extending a basis of the intersection proof
  87. Existence of an orthonormal basis statement
  88. Grassmann's formula statement
  89. Proof that a space is the direct sum of a finite-dimensional subspace and its orthogonal complement proof
  90. A space is the direct sum of a subspace and its orthogonal complement statement
  91. Orthogonal projection definition
  92. Proof of the formula for the dimension of the orthogonal complement proof
  93. Reflection definition
  94. Dimension of the orthogonal complement statement
  95. Proof of the Cartan–Dieudonné theorem by induction on the dimension proof

Where this leads

1 page builds on this one.

Further reading

  1. Web https://www.cis.upenn.edu/~cis6100/geombchap7.pdf, Lemma 7.1.3 and Theorem 7.2.1
About this page
TagAYYJXYNT01
URIm2://AYYJXYNT01
Kindproof
WritingW0 stub
MSC51F15, 15B10
ProvesCartan–Dieudonné theorem
UsesInner product, Linear isometry, A linear isometry preserves the inner product, Reflection, Orthogonal projection, Orthogonal complement, A space is the direct sum of a subspace and its orthogonal complement, Dimension of the orthogonal complement, A subspace of full dimension is the whole space
Changed4 November 2026

Markdown · JSON · Edit on GitHub

What this page needs for W1

  • an inventory row: the row of the inventory that the node accounts for (0 of 1)

The writing levels say how complete the text of a page is. They do not measure its truth.

What links here

No page links to this page yet.