Curvature measures the failure of covariant derivatives to commute
Fix the Riemann sign convention and connect invariant curvature to oriented geodesic deviation.
On this page
The question for today
What you’ll understand
- Define the Riemann tensor through a torsion-free commutator.
- Compute a flat-space cancellation despite nonzero connection coefficients.
- Apply the declared index ordering to a local tidal acceleration.
Christoffel coefficients can vanish at an event through a coordinate choice, so their values cannot alone measure intrinsic curvature. The Riemann tensor combines their derivatives and products into a coordinate-covariant object. It also controls the relative acceleration of neighboring geodesics. We will fix its sign and index order explicitly, because a formula for geodesic deviation can change apparent sign when either the curvature convention or the placement of the separation vector is changed.
The commutator removes coordinate artifacts
For the torsion-free Levi-Civita connection in a coordinate basis, the commutator of two covariant derivatives acting on a vector is algebraic in that vector. Expanding the two derivatives gives derivative-of-connection terms and quadratic connection terms. Their particular combination transforms as a tensor even though none of the connection coefficients does. The last two curvature indices are antisymmetric because they identify the two derivative directions. Reversing their order reverses the sign of the action.
Define R(X,Y)Z as ∇X∇YZ−∇Y∇XZ−∇[X,Y]Z. The vector Z occupies the second curvature slot in the component convention below, while X and Y occupy the last two. This matters in a noncoordinate frame, whose commutator may not vanish. Torsion would also modify the simple coordinate commutator formula. Throughout this module the metric signature is −+++ and the connection is torsion free and metric compatible; alternative curvature conventions in other references must be translated before comparing components.
A nonzero connection can still give zero Riemann tensor
In the flat polar plane, Γʳφφ=−r and Γφrφ=1/r. A typical curvature component contains a radial derivative −1 and a quadratic-connection contribution +1, which cancel exactly. Keeping only derivative terms would incorrectly label the plane curved. At the polar origin the coordinate formula cannot be evaluated by inserting r=0, because that point is outside this chart. Cartesian coordinates show regular flat geometry through the origin. The tensor criterion is vanishing curvature on regular patches, not small or finite values of every individual polar coefficient.
On a round two-sphere the cancellation instead leaves positive curvature in our convention, with orthonormal sectional curvature 1/a² for radius a. Parallel transport around a sufficiently small loop then depends on its oriented area. The direction of a loop matters, and reversing it reverses the leading holonomy. Infinitesimal holonomy is local curvature information. Noncontractible loops can additionally detect global identifications, so flatness alone need not establish path independence around every loop of an entire spacetime.
Relative acceleration survives a locally inertial frame
A freely falling observer can set local connection coefficients to zero at one event, eliminating the coordinate appearance of a gravitational acceleration there. Neighboring freely falling trajectories can still converge or separate. Geodesic deviation measures that relative effect to first order in their separation. Positive Rⁱ₀ⱼ₀ in a chosen spatial eigen-direction produces focusing under the displayed sign convention. The curvature must be evaluated along the reference trajectory, and treating it as constant is an additional short-time or special-geometry approximation.
The oscillator lab illustrates one constant focusing component after normalization. Reset and set both controls to damping ζ=0 and frequency ω₀=1; after one Step to t=0.5 s predict x=cos0.5≈0.878. Reset and set ζ=0 and ω₀=2; one Step predicts cos1≈0.540. Both start with displacement one and velocity zero. This matches the mathematical form of a constant positive tidal eigenvalue after assigning suitable units. It contains no spacetime metric, tensor components, or changing curvature, and should not be treated as a general geodesic-deviation solver.
Work it through
Use the declared tidal sign
In a parallel orthonormal rest frame let Rˣ₀ₓ₀=2.0×10⁻²³ m⁻² and an initially stationary neighboring geodesic have separation ξˣ=10 m. Take c=3.0×10⁸ m/s. Estimate the initial relative acceleration and the separation change after 100 s if the curvature and leading acceleration can be held constant.
The ordering in the rest-frame equation gives aˣ=−c²Rˣ₀ₓ₀ξˣ. The sign is negative for a positive separation and positive stated tidal component.
Its magnitude is (9.0×10¹⁶)×(2.0×10⁻²³)×10=1.8×10⁻⁵ m/s². This is a relative acceleration toward the reference trajectory.
The leading displacement change is (1/2)aˣt²=−0.5×1.8×10⁻⁵×10000=−0.090 m. This is small compared with the original ten metres, supporting the short-time estimate if the curvature also remains approximately constant.
The initial relative acceleration is −1.8×10⁻⁵ m/s² and the leading separation change after 100 s is −0.090 m, indicating focusing.
The living notebook
Explore the model
Interactive lab · oscillator
Compare two constant focusing rates
Reset and set BOTH controls to ζ=0,ω₀=1, then Step once to t=0.5 s: predict x≈0.878. Reset and set ζ=0,ω₀=2, then Step once: predict x≈0.540. The fixed initial data are x=1,v=0. This illustrates a normalized constant-coefficient deviation equation, not a metric solver.
- Displacement x(t)
- 1
- Velocity v(t)
- 0
- Regime
- underdamped
Unit mass, x(0) = 1 and v(0) = 0. The analytic solution includes underdamped, critically damped and overdamped cases; time is measured in seconds. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.
Look forA positive constant focusing eigenvalue produces oscillatory relative separation in this local linear model.
Pause & reason
Check your understanding
Your turn
Try it yourself
For the flat polar metric, evaluate Rʳφrφ from Γʳφφ=−r and Γφrφ=1/r on r>0. Show the derivative and quadratic terms separately and explain why substituting r=0 into the connection formula is inappropriate.
Take a moment to make a prediction or work it out on paper. Open a hint when you need one.
Give me a hint
The surviving expression is ∂rΓʳφφ−ΓʳφφΓφrφ. The two terms have the same dimension.
Compare with the solution
The derivative term is −1. The quadratic term is −(−r)(1/r)=+1, so Rʳφrφ=0 on the regular polar patch. The connection coefficients are nonzero but curvature cancels. At r=0 the polar chart and basis fail, so its divergent coefficient cannot be used to infer a curvature singularity; Cartesian coordinates cover that regular flat point.
Keep the idea
Take it with you
- The Riemann sign and index ordering must be fixed before a tidal calculation.
- Derivative and quadratic connection terms combine into invariant curvature.
- Geodesic deviation detects relative acceleration that a local inertial frame cannot remove.
Read further
Original sources and deeper explanations to follow at your own pace.
Optional progress, saved in this browser when available. Revisit any lesson at any time.
Lesson not marked complete.