|  Metric Tensors | 
On a Vector Space
  $ As inner product: A metric
    tensor on a vector space X is a symmetric bilinear map g:
    X × X → \(\mathbb R\) (rank-2 covariant tensor);
    Usually required to be non-degenerate, and in fact (except for spacetime metrics)
    positive-definite.
  $ As mapping: An isomorphism
    g: X → X* (index raising/lowering), by
    v \(\mapsto\)g(v, ).
  $ As orthonormal frame:
    A choice of basis for X, up to gauge.
  * Relationships: Any
    such metric induces a norm by ||x||:= g(x,
    x)1/2, and a distance by
d(x, y):= g(y−x, y−x)1/2 = || y−x || .
  * Generalizations: Degenerate and/or non-positive metrics.
  * Remark: If the signature
    has k minus signs, g ∈ GL(n, \(\mathbb R\))
    / SO(k, n−k).
  @ Generalizations: Fernández AACA(01)mp/02,
    AACA(01)mp/02 [metric "extensor"];
    Hammond IJMPD(13) [non-symmetric, and spin].
  > Online resources:
    see MathWorld page;
    Wikipedia page.
Metric Tensor Field on a Manifold > s.a. connection;
  curvature.
  * Idea: Globally, one of the
    ways in which one specifies the geometry of a differentiable manifold; Locally,
    a structure which gives all line elements at p congruent to any given
    line element at q; It implies conformal,
    projective and affine structures.
  $ Defs: A smooth assignment of a
    metric tensor on each TpM (inner
    product/mapping); Equivalent to a choice of orthonormal frame at each p ∈
    M, up to gauge.
  * Question: If (M, d)
    is such that d is a smooth function of 2 variables, can we define a metric
    g on M?
  @ References: Mendez a1507
      [analysis based on Takagi's factorization of the metric tensor].
  > Types and examples:
    see 2D, 3D, 4D
    manifolds; lorentzian and riemannian
    geometry; types of metrics.
  > Related topics: see hilbert space;
    metric decomposition, matching and
    perturbations.
Space of Metrics on a Manifold
  > see lorentzian geometries;
  riemannian geometries.
  @ General references: Demmel & Nink PRD(15)-a1506 [connections and geodesics].
  @ Characterization of metrics: Hervik CQG(11)-a1107 [ε-property];
    > s.a. distance between metrics.
Tensor Products of Metrics
  $ Def: Obtained by
    (M1, h1;
    M2, h2)
    \(\mapsto\) (M1
    × M2,
    π1* h1
    ⊗ π2*
    h2) .
  * Line element:
    Of the form ds2
    = gab(x)
    dxa
    dxb
    + hij(y)
    dyi
    dyj.
  * Connection:
    Γabc
    = same as those of gab;
    Γijk
    = same as those of hij;
    All Γs with mixed indices vanish.
  * Curvature:
    Rabcd
    and Rab
    = those of gab;
    R ijkl
    and Rij
    = those of hij; All mixed Rs vanish;
    R = gab
    Rab
    + hij
    Rij.
  * Warped Product: Obtained by
(M1, h1; M2, h2) \(\mapsto\) (M1 × M2, π1* h1 ⊗ exp{2θ} π2* h2) .
@ Warped product: Choi JMP(00)mp/02.
"No metric, No nothing" – J. Stachel
 main page
  – abbreviations
  – journals – comments
  – other sites – acknowledgements
  send feedback and suggestions to bombelli at olemiss.edu – modified 13 feb 2021