Curvature is tensorial: Difference between revisions

From Diffgeom
 
(13 intermediate revisions by the same user not shown)
Line 2: Line 2:
==Statement==
==Statement==


Let <math>\nabla</math> be a [[connection]] on a [[vector bundle]] <math>E</math> over a [[differential manifold]] <math>M</math>. The '''Riemann curvature tensor''' of <math>\nabla</math> is given as a map <math>\Gamma(TM) \otimes \Gamma(TM) \otimes \Gamma(E) \to \Gamma(E)</math> defined by:
Let <math>\nabla</math> be a [[connection]] on a [[vector bundle]] <math>E</math> over a [[differential manifold]] <math>M</math>. The [[fact about::Riemann curvature tensor]] of <math>\nabla</math> is given as a map <math>\Gamma(TM) \otimes \Gamma(TM) \otimes \Gamma(E) \to \Gamma(E)</math> defined by:


<math>R(X,Y)Z = \nabla_X\nabla_Y Z - \nabla_Y \nabla_X Z - \nabla_{[X,Y]} Z</math>
<math>R(X,Y)Z = \nabla_X\nabla_Y Z - \nabla_Y \nabla_X Z - \nabla_{[X,Y]} Z</math>


We claim that <math>R</math> is a [[tensorial map]] in each of the variables <math>X,Y,Z</math>.
We claim that <math>R</math> is a [[fact about::tensorial map]] in each of the variables <math>X,Y,Z</math>.


==Facts used==
==Related facts==


* [[Leibniz rule for derivations]]: This states that for a vector field <math>X</math> and functions <math>f,g</math>, we have:
* [[Curvature is antisymmetric in first two variables]]
* [[Curvature is antisymmetric in last two variables]]
* [[Curvature is symmetric in the pairs of first and last two variables]]


<math>X(fg) = (Xf)(g) + f(Xg)</math>
==Facts used==
 
* [[Corollary of Leibniz rule for Lie bracket]]: This states that for a function <math>f</math> and vector fields <math>X,Y</math>:


<math>f[X,Y] = [fX,Y] + (Yf)X</math>
{| class="sortable" border="1"
 
! Fact no. !! Name !! Statement with symbols
<math>f[X,Y] = [X,fY] - (Xf)Y</math>
|-
 
| 1 || Any connection is <math>C^\infty</math>-linear in its subscript argument || <math>\nabla_{fA} = f\nabla_A</math> for any <math>C^\infty</math>-function <math>f</math> and vector field <math>A</math>.
* The Leibniz rule axiom that's part of the definition of a [[connection]], namely:
|-
 
| 2 || The Leibniz-like axiom that is part of the definition of a connection || For a function <math>f</math> and vector fields <math>A,B</math>, and a connection <math>\nabla</math>, we have <math>\nabla_A(fB) = (Af)(B) + f\nabla_A(B)</math>
<math>\nabla_X(fZ) = (Xf)(Z) + f\nabla_X(Z)</math>
|-
| 3 || [[uses::Corollary of Leibniz rule for Lie bracket]] (in turn follows from [[uses::Leibniz rule for derivations]]|| For a function <math>f</math> and vector fields <math>X,Y</math>:
<br><math>\! f[X,Y] = [fX,Y] + (Yf)X</math><br><math>\! f[X,Y] = [X,fY] - (Xf)Y</math>
|}


==Proof==
==Proof==
Line 32: Line 35:
===Tensoriality in the first variable===
===Tensoriality in the first variable===


Let <math>f:M \to \R</math> be a scalar function. We will show that:
'''Given''': <math>f:M \to \R</math> is a <math>C^\infty</math>-function.


<math>R(fX,Y) = f R(X,Y)</math>
'''To prove''': <math>\! R(fX,Y) = f R(X,Y)</math>, or more explicitly, <math>\! \nabla_{fX}\nabla_Y - \nabla_Y \nabla_{fX} - \nabla_{[fX,Y]} = f(\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]}</math>


We start out with the left side:
We start out with the left side:
Line 40: Line 43:
<math>\nabla_{fX}\nabla_Y - \nabla_Y \nabla_{fX} - \nabla_{[fX,Y]}</math>
<math>\nabla_{fX}\nabla_Y - \nabla_Y \nabla_{fX} - \nabla_{[fX,Y]}</math>


Now by the definition of a [[connection]], <math>\nabla</math> is <math>C^\infty</math>-linear in its subscript argument. Thus, the above expression can be written as:
Each step below is obtained from the previous one via some manipulation explained along side.
 
<math>f\nabla_X\nabla_Y - \nabla_Y (f \nabla_X) - \nabla_{[fX,Y]}</math>
 
Now applying the Leibniz rule for connections, we get:
 
<math>f\nabla_X\nabla_Y - (Yf)\nabla_X - f \nabla_Y\nabla_X - \nabla_{[fX,Y]}</math>
 
We can rewrite <math>(Yf)\nabla_X = \nabla_{(Yf)X}</math> and we then get:


<math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X) - \nabla_{(Yf)X + [fX,Y]}</math>
{| class="sortable" border="1"
 
! Step no. !! Current status of left side !! Facts/properties used !! Specific rewrites
By the corollary stated above, we have:
|-
 
| 1 || <math>f\nabla_X\nabla_Y - \nabla_Y (f \nabla_X) - \nabla_{[fX,Y]}</math> || Fact (1): <math>\nabla</math> is <math>C^\infty</math>-linear in its subscript argument. || <math>\nabla_{fX} \to f\nabla_X</math>
<math>(Yf)X + [fX,Y] = f[X,Y]</math>
|-
 
| 2 || <math>f\nabla_X\nabla_Y - (Yf)\nabla_X - f \nabla_Y\nabla_X - \nabla_{[fX,Y]}</math> || Fact (2) || <math>\nabla_Y(f \nabla_X) \to (Yf)\nabla_X + f\nabla_Y\nabla_X</math>. To understand this more clearly imagine an input <math>Z</math> to the whole expression, so that the rewrite becomes <math>\nabla_Y(f \nabla_X(Z)) \to (Yf)\nabla_X(Z) + f\nabla_Y\nabla_X(Z)</math>. In the notation of fact (3), <math>A = Y</math>, <math>f = f</math>, and <math>B = \nabla_X(Z)</math>.
which, substituted back, gives:
|-
 
| 3 || <math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X) - \nabla_{(Yf)X} - \nabla_{[fX,Y]}</math> || Fact (1) || <math>(Yf)\nabla_X \to \nabla_{(Yf)X}</math>
<math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]})</math>
|-
| 4 || <math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X) - \nabla_{(Yf)X + [fX,Y]}</math> || <math>\nabla</math> is additive in its subscript argument || <math>\nabla_{(Yf)X} + \nabla_{[fX,Y]} = \nabla_{(Yf)X + [fX,Y]}</math>
|-
| 5 || <math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X) - \nabla_{f[X,Y]}</math> || Fact (3) || <math>[fX,Y] + (Yf)X \to f[X,Y]</math>
|-
| 6 || <math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]})</math> || Fact (1) || <math>\nabla_{f[X,Y]} \to f\nabla_{[X,Y]}</math>
|}


===Tensoriality in the second variable===
===Tensoriality in the second variable===


Let <math>f:M \to \R</math> be a scalar function. We will show that:
'''Given''': <math>f:M \to \R</math> is a <math>C^\infty</math>-function.


<math>R(X,fY) = f R(X,Y)</math>
'''To prove''': <math>\! R(X,fY) = f R(X,Y)</math>, or more explicitly, <math>\nabla_X\nabla_{fY} - \nabla_{fY}\nabla_X - \nabla_{[X,fY]} = f (\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]}</math>.


We start out with the left side:
We start out with the left side:
Line 70: Line 71:
<math>\nabla_X\nabla_{fY} - \nabla_{fY}\nabla_X - \nabla_{[X,fY]}</math>
<math>\nabla_X\nabla_{fY} - \nabla_{fY}\nabla_X - \nabla_{[X,fY]}</math>


Applying the Leibniz rule and the property of a connection being <math>C^\infty</math> in its subscript variable yields:
Each step below is obtained from the previous one via some manipulation explained along side.


<math>(Xf)\nabla_Y + f(\nabla_X\nabla_Y - \nabla_Y\nabla_X) - \nabla_{[X,fY]}</math>
{| class="sortable" border="1"
 
! Step no. !! Current status of left side !! Facts/properties used !! Specific rewrites
which simplifies to:
|-
 
| 1 || <math>\nabla_X(f\nabla_Y) - f\nabla_Y\nabla_X - \nabla_{[X,fY]}</math> || Fact (1) || <math>\nabla_{fY} \to f\nabla_Y</math>.
<math>f(\nabla_X\nabla_y - \nabla_Y\nabla_X) - \nabla_{[X,fY] - (Xf)Y}</math>
|-
 
| 2 || <math>(Xf)\nabla_Y + f(\nabla_X\nabla_Y) - f\nabla_Y\nabla_X - \nabla_{[X,fY]}</math> || Fact (2) || <math>\nabla_X(f\nabla_Y) \to (Xf)\nabla_Y + f(\nabla_X\nabla_Y)</math>. To make this more concrete, imagine an input <math>Z</math>. Then, the rewrite becomes <math>\nabla_X(f\nabla_Y(Z)) \to (Xf)\nabla_Y(X) + f(\nabla_X\nabla_Y(Z))</math>. This comes setting <math>A = X</math>, <math>f = f</math>, <math>B = \nabla_YZ</math> in Fact (3).
We now use the corollary stated above:
|-
 
| 3 || <math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X) - \nabla_{[X,fY]} + \nabla_{(Xf)Y}</math> || Fact (1) || <math>(Xf)\nabla_Y \to \nabla_{(Xf)Y}</math>
<math>f[X,Y] = [X,fY] - (Xf)Y</math>
|-
 
| 4 || <math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X) - \nabla_{[X,fY] - (Xf)Y}</math> || <math>\nabla</math> is additive in its subscript argument. || <math>\nabla_{[X,fY]} - \nabla_{(Xf)Y} \to \nabla_{[X,fY] - (Xf)Y}</math>.
substituting this gives:
|-
 
| 5 || <math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X) - \nabla_{f[X,Y]}</math> || Fact (3) || <math>[X,fY] - (Xf)Y \to f[X,Y]</math>
<math>f (\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]}</math>
|-
 
| 6 || <math>f(\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]})</math> || Fact (1) || <math>\nabla_{f[X,Y]} \to f\nabla_{[X,Y]}</math>
which is <math>f R(X,Y)</math>
|}


===Tensoriality in the third variable===
===Tensoriality in the third variable===


Let <math>f: M \to \R</math> be a scalar function. We will show that:
'''Given''': A <math>C^\infty</math>-function <math>f:M \to \R</math>.


<math>R(X,Y) (fZ) = f R(X,Y) Z</math>
'''To prove''': <math>\! R(X,Y) (fZ) = f R(X,Y) Z</math>. More explicitly, <math>\! \nabla_X\nabla_Y(fZ) - \nabla_Y\nabla_X(fZ) - \nabla_{[X,Y]}(fZ)  = f (\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]})Z + ((XY - YX - [X,Y])f)Z</math>.


We start out with the left side:
We start out with the left side:
Line 98: Line 99:
<math>\nabla_X\nabla_Y(fZ) - \nabla_Y\nabla_X(fZ) - \nabla_{[X,Y]}(fZ)</math>
<math>\nabla_X\nabla_Y(fZ) - \nabla_Y\nabla_X(fZ) - \nabla_{[X,Y]}(fZ)</math>


Now we apply the Leibniz rule for connnections on each term:
Each step below is obtained from the previous one via some manipulation explained along side.
 
<math>\nabla_X( (Yf)(Z) + f \nabla_YZ) - \nabla_Y ((Xf)Z + f \nabla_XZ) - f \nabla_{[X,Y]}Z - ([X,Y]f) Z</math>
 
We again apply the Leibniz rule to the first two term groups:
 
<math>(XYf)(Z) + (Yf) \nabla_XZ + (Xf) \nabla_YZ + f \nabla_X\nabla_YZ - (YXf)Z - (Xf) \nabla_YZ - (Yf) \nabla_XZ -f \nabla_Y\nabla_XZ - f \nabla_{[X,Y]}Z - ([X,Y]f) Z</math>
 
After cancellations we are left with the following six terms:
 
<math>f (\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]})Z + ((XY - YX - [X,Y])f)Z</math>
 
But since <math>[X,Y] = XY - YX</math>, the last three terms vanish, and we are left with:


<math>f R(X,Y)Z</math>
{| class="sortable" border="1"
! Step no. !! Current status of left side !! Facts/properties used !! Specific rewrites
|-
| 1 || <math>\! \nabla_X( (Yf)(Z) + f \nabla_YZ) - \nabla_Y ((Xf)Z + f \nabla_XZ) - f \nabla_{[X,Y]}Z - ([X,Y]f) Z</math> || Fact (2) || <math>\nabla_Y(fZ) \to (Yf)(Z) + f\nabla_YZ</math> and <math>\nabla_X(fZ) \to (Xf)Z + f\nabla_XZ</math>
|-
| 2 || <math>\! (XYf)(Z) + (Yf) \nabla_XZ + (Xf) \nabla_YZ + f \nabla_X\nabla_YZ - (YXf)Z - (Xf) \nabla_YZ - (Yf) \nabla_XZ -f \nabla_Y\nabla_XZ - f \nabla_{[X,Y]}Z - ([X,Y]f) Z</math> || Fact (2) || <math>\nabla_X((Yf)Z) \to X((Yf)Z) + (Yf)\nabla_XZ</math>, etc.
|-
| 3 || <math>f (\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]})Z + ((XY - YX - [X,Y])f)Z</math> || -- || cancellations
|-
| 4 || <math>f (\nabla_X\nabla_Y - \nabla_Y\nabla_X - \nabla_{[X,Y]})Z + ((XY - YX - [X,Y])f)Z</math>|| use <math>[X,Y] = XY - YX</math>, definition || cancellation
|}

Latest revision as of 17:36, 6 January 2012

This article gives the statement, and possibly proof, that a map constructed in a certain way is tensorial
View other such statements

Statement

Let be a connection on a vector bundle E over a differential manifold M. The Riemann curvature tensor of is given as a map Γ(TM)Γ(TM)Γ(E)Γ(E) defined by:

R(X,Y)Z=XYZYXZ[X,Y]Z

We claim that R is a tensorial map in each of the variables X,Y,Z.

Related facts

Facts used

Fact no. Name Statement with symbols
1 Any connection is C-linear in its subscript argument fA=fA for any C-function f and vector field A.
2 The Leibniz-like axiom that is part of the definition of a connection For a function f and vector fields A,B, and a connection , we have A(fB)=(Af)(B)+fA(B)
3 Corollary of Leibniz rule for Lie bracket (in turn follows from Leibniz rule for derivations For a function f and vector fields X,Y:


f[X,Y]=[fX,Y]+(Yf)X
f[X,Y]=[X,fY](Xf)Y

Proof

To prove tensoriality in a variable, it suffices to show C-linearity in that variable. This is because linearity in C-functions guarantees linearity in a function that is 1 at exactly one point, and zero at others.

The proofs for X and Y are analogous, and rely on manipulation of the Lie bracket [fX,Y] and the property of a connection being C in the subscript vector. These proofs do not involve any explicit use of Z. The proof for Z relies simply on repeated application of the product rule, and the fact that XYYX=[X,Y].

Tensoriality in the first variable

Given: f:MR is a C-function.

To prove: R(fX,Y)=fR(X,Y), or more explicitly, fXYYfX[fX,Y]=f(XYYX[X,Y]

We start out with the left side:

fXYYfX[fX,Y]

Each step below is obtained from the previous one via some manipulation explained along side.

Step no. Current status of left side Facts/properties used Specific rewrites
1 fXYY(fX)[fX,Y] Fact (1): is C-linear in its subscript argument. fXfX
2 fXY(Yf)XfYX[fX,Y] Fact (2) Y(fX)(Yf)X+fYX. To understand this more clearly imagine an input Z to the whole expression, so that the rewrite becomes Y(fX(Z))(Yf)X(Z)+fYX(Z). In the notation of fact (3), A=Y, f=f, and B=X(Z).
3 f(XYYX)(Yf)X[fX,Y] Fact (1) (Yf)X(Yf)X
4 f(XYYX)(Yf)X+[fX,Y] is additive in its subscript argument (Yf)X+[fX,Y]=(Yf)X+[fX,Y]
5 f(XYYX)f[X,Y] Fact (3) [fX,Y]+(Yf)Xf[X,Y]
6 f(XYYX[X,Y]) Fact (1) f[X,Y]f[X,Y]

Tensoriality in the second variable

Given: f:MR is a C-function.

To prove: R(X,fY)=fR(X,Y), or more explicitly, XfYfYX[X,fY]=f(XYYX[X,Y].

We start out with the left side:

XfYfYX[X,fY]

Each step below is obtained from the previous one via some manipulation explained along side.

Step no. Current status of left side Facts/properties used Specific rewrites
1 X(fY)fYX[X,fY] Fact (1) fYfY.
2 (Xf)Y+f(XY)fYX[X,fY] Fact (2) X(fY)(Xf)Y+f(XY). To make this more concrete, imagine an input Z. Then, the rewrite becomes X(fY(Z))(Xf)Y(X)+f(XY(Z)). This comes setting A=X, f=f, B=YZ in Fact (3).
3 f(XYYX)[X,fY]+(Xf)Y Fact (1) (Xf)Y(Xf)Y
4 f(XYYX)[X,fY](Xf)Y is additive in its subscript argument. [X,fY](Xf)Y[X,fY](Xf)Y.
5 f(XYYX)f[X,Y] Fact (3) [X,fY](Xf)Yf[X,Y]
6 f(XYYX[X,Y]) Fact (1) f[X,Y]f[X,Y]

Tensoriality in the third variable

Given: A C-function f:MR.

To prove: R(X,Y)(fZ)=fR(X,Y)Z. More explicitly, XY(fZ)YX(fZ)[X,Y](fZ)=f(XYYX[X,Y])Z+((XYYX[X,Y])f)Z.

We start out with the left side:

XY(fZ)YX(fZ)[X,Y](fZ)

Each step below is obtained from the previous one via some manipulation explained along side.

Step no. Current status of left side Facts/properties used Specific rewrites
1 X((Yf)(Z)+fYZ)Y((Xf)Z+fXZ)f[X,Y]Z([X,Y]f)Z Fact (2) Y(fZ)(Yf)(Z)+fYZ and X(fZ)(Xf)Z+fXZ
2 (XYf)(Z)+(Yf)XZ+(Xf)YZ+fXYZ(YXf)Z(Xf)YZ(Yf)XZfYXZf[X,Y]Z([X,Y]f)Z Fact (2) X((Yf)Z)X((Yf)Z)+(Yf)XZ, etc.
3 f(XYYX[X,Y])Z+((XYYX[X,Y])f)Z -- cancellations
4 f(XYYX[X,Y])Z+((XYYX[X,Y])f)Z use [X,Y]=XYYX, definition cancellation