Geometric Origin of Fictitious Forces via Christoffel Symbols

By Ram Prakash | July 2026 | Category: Mathematical Physics

Figure 0: Geometric visualization of moving frame coordinate bases.

One often doesn't realize the importance of differential geometry while dealing with classical mechanics. In this blog I want to point out tensorial way of expressing Newton's Second law, and emphasize on importance of including Christoffel symbols while defining acceleration to deal with change of basis frames in curved space.

In classical dynamics, observers in rotating or accelerating reference frames experience "fictitious forces" like the centrifugal and Coriolis forces. In Cartesian coords, Newton's second law is clean and straightforward. However, as soon as we transform to curvilinear coordinate grids (like polar coordinates) or transition to accelerating observers, correction terms arise.

When viewed through the lens of differential geometry, these fictitious forces are purely geometric phenomena. They correspond to the coordinate-dependent components of the metric connection, mathematically encoded in the Christoffel symbols. Below, we derive these forces from first principles using the Lagrangian formulation of a classical particle in a potential field.

Additionally, we present an alternative physical derivation of the Christoffel symbols and the general covariant derivative in the Appendix. This approach relies on the Equivalence Principle, starting from the physical insight that in a locally flat (inertial) space, parallel transport of a vector is trivial, and derivatives reduce to standard directional derivatives. Transforming this flat-space derivative into a curved coordinate frame naturally yields the connection terms.

1. Lagrangian of a Classical Particle

Consider a particle of mass \(m\) moving in a space characterized by a metric tensor \(g_{\mu\nu}\). The action \(S\) for this trajectory is:

\[ S = \int_{t_1}^{t_2} dt \, L \]

where the classical Lagrangian \(L\) for a particle subject to a potential \(V(x)\) is defined as:

\[ L = \frac{m}{2} g_{\mu\nu} \dot{x}^\mu \dot{x}^\nu - V(x) \]

Here, the coordinate derivatives are with respect to the time parameter \(t\) (i.e., \(\dot{x}^\mu = \frac{dx^\mu}{dt}\)). According to the Euler-Lagrange equations of motion:

\[ \frac{d}{dt} \left( \frac{\partial L}{\partial \dot{x}^\mu} \right) = \frac{\partial L}{\partial x^\mu} \]

Let's evaluate each term carefully. First, the momentum conjugate to \(x^\mu\) is:

\[ \frac{\partial L}{\partial \dot{x}^\mu} = m g_{\mu\nu} \dot{x}^\nu \]

Taking its total time derivative:

\[ \frac{d}{dt} \left( m g_{\mu\nu} \dot{x}^\nu \right) = m g_{\mu\nu} \ddot{x}^\nu + m \left( \partial_\alpha g_{\mu\nu} \right) \dot{x}^\alpha \dot{x}^\nu \]

Next, the right-hand side of the Euler-Lagrange equation yields:

\[ \frac{\partial L}{\partial x^\mu} = \frac{m}{2} \left( \partial_\mu g_{\alpha\nu} \right) \dot{x}^\alpha \dot{x}^\nu - \partial_\mu V \]

Equating the two parts, we get:

\[ m g_{\mu\nu} \ddot{x}^\nu + m g_{\mu\nu,\alpha} \dot{x}^\alpha \dot{x}^\nu - \frac{m}{2} g_{\alpha\nu,\mu} \dot{x}^\alpha \dot{x}^\nu = - \partial_\mu V \]

where we write \(g_{\alpha\beta,\mu} = \partial_\mu g_{\alpha\beta}\) for the partial derivatives of the metric. Symmetrizing the second term (since it is contracted with symmetric velocity components \(\dot{x}^\alpha \dot{x}^\nu\)), we rearrange this into:

\[ g_{\mu\nu} \ddot{x}^\nu + \frac{1}{2} \left[ g_{\mu\nu,\alpha} + g_{\mu\alpha,\nu} - g_{\alpha\nu,\mu} \right] \dot{x}^\alpha \dot{x}^\nu = -\frac{1}{m} \partial_\mu V \]

Multiplying both sides by the inverse metric \(g^{\gamma\mu}\), we isolate the coordinate acceleration \(\ddot{x}^\gamma\):

\[ \ddot{x}^\gamma + g^{\gamma\mu} \frac{1}{2} \left[ g_{\mu\nu,\alpha} + g_{\mu\alpha,\nu} - g_{\alpha\nu,\mu} \right] \dot{x}^\alpha \dot{x}^\nu = -\frac{1}{m} g^{\gamma\mu} \partial_\mu V \]

The term in brackets multiplied by the inverse metric is exactly the definition of the Christoffel symbol of the second kind \(\Gamma^\gamma_{\alpha\nu}\):

Definition: Christoffel Symbol of the Second Kind

\[ \Gamma^\gamma_{\alpha\nu} = \frac{1}{2} g^{\gamma\mu} \left( g_{\mu\nu,\alpha} + g_{\mu\alpha,\nu} - g_{\alpha\nu,\mu} \right) \]

Physically, the Christoffel symbols describe how coordinate basis vectors change orientation as a function of position on the manifold.

2. Equation of Motion

Using our definition of the Christoffel symbols, the equation of motion simplifies to:

\[ \ddot{x}^\gamma + \Gamma^\gamma_{\alpha\nu} \dot{x}^\alpha \dot{x}^\nu = -\frac{1}{m} g^{\gamma\mu} \partial_\mu V = -\frac{1}{m} \partial^\gamma V \]

Note that this action is not invariant under general time reparameterization, the parameter \(t\) is chosen as the Newtonian time. However, writing the equation of motion in terms of the covariant derivative along the curve \(c(t)\) yields a remarkably elegant, geometric statement:

\[ \nabla_{c'(t)} c'(t) = -\frac{1}{m} g^{\gamma\mu} \partial_\mu V \]

This tells us that the covariant derivative of velocity of the curve, along the curve (which is the true coordinate-free geometric acceleration) is driven by the potential gradient. The actual coordinate acceleration \(\ddot{x}^\gamma\) is not equal to the true acceleration. Instead, the true acceleration is the coordinate acceleration plus the geometric connection term \(\Gamma^\gamma_{\alpha\nu} \dot{x}^\alpha \dot{x}^\nu\). When no potential is present (\(V = 0\)), this equation reduces to the geodesic equation:

\[ \nabla_{c'(t)} c'(t) = 0 \implies \ddot{x}^\gamma + \Gamma^\gamma_{\alpha\nu} \dot{x}^\alpha \dot{x}^\nu = 0 \]

In a flat space using curvilinear coordinates, or in a curved manifold under gravity, a free particle simply traces out a straight line (geodesic). What we perceive as forces acting on the particle are just geometric corrections needed to account for the bending of our coordinate axes.

3. Example: Particle on a 2D Plane

To make this concrete, let us evaluate the equations of motion for a particle on a flat 2D plane using polar coordinates \((r, \theta)\) instead of Cartesian ones.

_r ê_θ

Figure 1: Polar coordinates with unit basis vectors.

The metric tensor \(g_{\mu\nu}\) and its inverse \(g^{\alpha\beta}\) for polar coordinates are:

\[ g_{\mu\nu} = \begin{pmatrix} 1 & 0 \\ 0 & r^2 \end{pmatrix}, \quad g^{\alpha\beta} = \begin{pmatrix} 1 & 0 \\ 0 & \frac{1}{r^2} \end{pmatrix} \]

Calculating the Christoffel symbols using their definition, the only non-zero components are:

\[ \Gamma^r_{\theta\theta} = -r, \quad \Gamma^\theta_{r\theta} = \Gamma^\theta_{\theta r} = \frac{1}{r} \]

Substituting these values into the geodesic equation \(\ddot{x}^\gamma + \Gamma^\gamma_{\alpha\nu} \dot{x}^\alpha \dot{x}^\nu = 0\):

For \(\gamma = r\):
\[ \ddot{r} + \Gamma^r_{\theta\theta} \dot{\theta}^2 = 0 \implies \ddot{r} - r \dot{\theta}^2 = 0 \]
For \(\gamma = \theta\):
\[ \ddot{\theta} + 2\Gamma^\theta_{r\theta} \dot{r}\dot{\theta} = 0 \implies \ddot{\theta} + \frac{2}{r} \dot{r}\dot{\theta} = 0 \]

Let's introduce the potential terms back into the equations: \(F_r = -\partial_r V\) and \(F_\theta = -\partial_\theta V\). Multiplying the radial equation by mass \(m\), and the angular equation by \(mr^2\), we obtain:

                    \[ m \ddot{r} = F_r + m r \dot{\theta}^2 \quad \text{(Radial Component)} \]
                    \[ m r^2 \ddot{\theta} + 2 m r \dot{r} \dot{\theta} = F_\theta \implies m r \ddot{\theta} = \frac{F_\theta}{r} - 2 m \dot{r} \dot{\theta} \quad \text{(Angular Component)} \]
                

Look at the resulting terms on the right-hand sides:

Centrifugal Force (\(m r \dot{\theta}^2\)): Appears as an apparent outward radial force because the grid lines of \(\theta\) curve.
Coriolis Force (\(-2 m \dot{r} \dot{\theta}\)): Appears in the angular direction to conserve angular momentum when the radial distance changes.

These forces arise naturally as the geometric corrections (encoded in Christoffel symbols) that reconcile straight-line motion with curved coordinate axes.

4. Christoffel Symbols as Coefficients of Basis Rotation

Why do these connection terms appear? Let's consider a vector field \(\vec{V} = \frac{\partial}{\partial \theta}\) representing a velocity field circulating around the origin in polar coordinates. Its coordinate components are \(V^r = 0\) and \(V^\theta = 1\).

Naively, if we look only at the components, one might say that because \(V^r\) and \(V^\theta\) are constant along circular paths, the rate of change is zero:

\[ \frac{dV^\mu}{dt} = 0 \]

However, this is physically incorrect. The vector field represents circular flow, which clearly undergoes centripetal acceleration in Cartesian coordinates.

Figure 2: Circulating vector field V.

The error lies in ignoring that the coordinate basis vectors themselves (like \(\vec{\partial}_\theta\)) rotate and change as the particle shifts. The true covariant rate of change of a vector field \(\vec{V}\) along a trajectory must use the product rule to account for the basis change:

\[ \frac{D V^\alpha}{Dt} = \frac{d V^\alpha}{dt} + V^\mu \left( \frac{d \vec{\partial}_\mu}{dt} \right)^\alpha \]

The rate of change of the coordinate basis vectors themselves is defined by the Christoffel symbols:

\[ \frac{d \vec{\partial}_\mu}{dt} = \Gamma^\alpha_{\mu\nu} \dot{x}^\nu \vec{\partial}_\alpha \]

Combining these, the component representation of the covariant derivative along the curve \(c(t)\) is:

                    \[ \left( \nabla_{c'(t)} V \right)^\mu = \frac{dV^\mu}{dt} + \Gamma^\mu_{\alpha\beta} V^\alpha \dot{x}^\beta \]
                

Physical & Geometric Interpretation: For a coordinate basis (which is tangent to the coordinate curves and may be orthogonal but is generally not orthonormal), the Christoffel symbol \(\Gamma^\mu_{\alpha\beta}\) represents the \(\mu\)-th component of the change of the \(\alpha\)-th coordinate basis vector along the \(\beta\)-th coordinate direction.

Example: Deriving Christoffel Symbols Geometrically by Considering Rotation of Basis Frames

Question: Derive the non-vanishing Christoffel symbols in polar coordinates \((r, \theta)\) purely geometrically, by analyzing how the coordinate tangent basis vectors change as we navigate along the coordinate lines.

Solution:

In terms of Cartesian basis vectors \(\hat{i}\) and \(\hat{j}\), the position vector of a point in polar coordinates is:

\[ \vec{r} = r \cos\theta \hat{i} + r \sin\theta \hat{j} \]

The coordinate tangent vectors (unnormalized) are defined by the partial derivatives of the position vector with respect to each coordinate parameter:

\[ \vec{\partial}_r = \frac{\partial \vec{r}}{\partial r} = \cos\theta \hat{i} + \sin\theta \hat{j} \] \[ \vec{\partial}_\theta = \frac{\partial \vec{r}}{\partial \theta} = -r \sin\theta \hat{i} + r \cos\theta \hat{j} \]

We now compute the spatial derivatives of these coordinate tangent vectors \(\vec{\partial}_r\) and \(\vec{\partial}_\theta\) with respect to the coordinates \(r\) and \(\theta\):

Derivatives with respect to \(r\) (moving radially):
Moving radially outward, the direction of \(\vec{\partial}_r\) does not change:
\[ \frac{\partial \vec{\partial}_r}{\partial r} = \frac{\partial}{\partial r} \left( \cos\theta \hat{i} + \sin\theta \hat{j} \right) = 0 \]
However, the angular tangent vector \(\vec{\partial}_\theta\) grows linearly with \(r\) without changing its direction:
\[ \frac{\partial \vec{\partial}_\theta}{\partial r} = \frac{\partial}{\partial r} \left( -r \sin\theta \hat{i} + r \cos\theta \hat{j} \right) = -\sin\theta \hat{i} + \cos\theta \hat{j} = \vec{e}_\theta = \frac{1}{r} \vec{\partial}_\theta \]
Derivatives with respect to \(\theta\) (moving angularly / rotating the frame):
Rotating the frame changes the direction of the radial tangent vector \(\vec{\partial}_r\), tilting it in the tangential direction:
\[ \frac{\partial \vec{\partial}_r}{\partial \theta} = \frac{\partial}{\partial \theta} \left( \cos\theta \hat{i} + \sin\theta \hat{j} \right) = -\sin\theta \hat{i} + \cos\theta \hat{j} = \vec{e}_\theta = \frac{1}{r} \vec{\partial}_\theta \]
Similarly, rotating the frame tilts the tangential tangent vector \(\vec{\partial}_\theta\) inward toward the negative radial direction:
\[ \frac{\partial \vec{\partial}_\theta}{\partial \theta} = \frac{\partial}{\partial \theta} \left( -r \sin\theta \hat{i} + r \cos\theta \hat{j} \right) = -r \cos\theta \hat{i} - r \sin\theta \hat{j} = -r \vec{e}_r = -r \vec{\partial}_r \]

By definition, the Christoffel symbols of the second kind \(\Gamma^\gamma_{\alpha\beta}\) are the expansion coefficients describing how basis vectors change relative to the coordinate basis frame:

\[ \frac{\partial \vec{\partial}_\alpha}{\partial x^\beta} = \Gamma^\gamma_{\alpha\beta} \vec{\partial}_\gamma \]

By matching terms from our spatial derivatives, we extract the non-vanishing coefficients:

• For \(\frac{\partial \vec{\partial}_r}{\partial r} = 0 \implies \Gamma^r_{rr} = 0, \quad \Gamma^\theta_{rr} = 0\)
• For \(\frac{\partial \vec{\partial}_r}{\partial \theta} = 0 \cdot \vec{\partial}_r + \frac{1}{r} \vec{\partial}_\theta \implies \Gamma^r_{r\theta} = 0, \quad \Gamma^\theta_{r\theta} = \frac{1}{r}\)
• By coordinate symmetry (\(\frac{\partial^2 \vec{r}}{\partial r \partial \theta} = \frac{\partial^2 \vec{r}}{\partial \theta \partial r}\)): \(\Gamma^r_{\theta r} = 0, \quad \Gamma^\theta_{\theta r} = \frac{1}{r}\)
• For \(\frac{\partial \vec{\partial}_\theta}{\partial \theta} = -r \vec{\partial}_r + 0 \cdot \vec{\partial}_\theta \implies \Gamma^r_{\theta\theta} = -r, \quad \Gamma^\theta_{\theta\theta} = 0\)

This completes the geometric derivation of the polar Christoffel symbols, showing that they represent the rate of rotation and scaling of the coordinate basis frames as we move across the manifold.

This result shows that Christoffel symbols act as the coefficients of basis rotation and scaling. They correct the standard derivative, accounting for the curving, scaling and rotation of the coordinate grid. Without them, we would misidentify rotating basis coordinates as non-accelerating, violating the core coordinate invariance of physical laws.

Appendix: Derivation of the Covariant Derivative

From a Locally Inertial Frame to a General Coordinate System

An alternative, deeply physical way to understand the covariant derivative and the origin of the Christoffel symbols is through the Equivalence Principle. The principle states that at any point in a curved spacetime (or under an arbitrary coordinate transformation), we can always construct a locally inertial frame (LIF) where the effects of gravity and acceleration vanish locally.

In this appendix, we show that the covariant derivative is simply the ordinary directional derivative in this locally inertial frame, which, when transformed back into our general coordinates, gives rise to the connection terms containing Christoffel symbols.

1. The Locally Inertial Frame (LIF)

Let \(x^\mu\) represent coordinates in the locally inertial frame at a given point on the manifold. In this frame, the metric is locally flat (approximating the Minkowski metric \(\eta_{\mu\nu}\)), and its first derivatives vanish at that point: \(\partial_\lambda \eta_{\mu\nu} = 0\).

For a vector field \(V\) defined along a curve parameterized by \(t\), the covariant derivative along the curve in this inertial frame reduces entirely to the standard ordinary derivative, as there is no coordinate bending or fictitious force acting locally:

\[ \nabla_{\frac{\partial}{\partial t}} V^\mu = \frac{dV^\mu}{dt} \tag{1} \]

2. Transformation to a General Frame

Now, let us transform this result into a general coordinate system with coordinates \(x'^\alpha\), where the metric is \(g_{\alpha\beta}\). The vector components in the two frames are related by the standard tensor transformation law:

\[ V^\nu = \frac{\partial x^\nu}{\partial x'^\alpha} V'^\alpha \tag{2} \]

To evaluate how the covariant derivative transforms, we note that it is a tensor, so its components in the new coordinate system, denoted by \(\left(\nabla_{\frac{\partial}{\partial t}} V^\mu\right)'\), are related to the LIF components by the tensor transformation rule:

\[ \left(\nabla_{\frac{\partial}{\partial t}} V^\mu\right)' = \frac{\partial x'^\mu}{\partial x^\nu} \left( \frac{dV^\nu}{dt} \right) \]

Substituting the transformation of the vector (Equation 2) into the derivative term, we get:

\[ \left(\nabla_{\frac{\partial}{\partial t}} V^\mu\right)' = \frac{\partial x'^\mu}{\partial x^\nu} \frac{d}{dt} \left( \frac{\partial x^\nu}{\partial x'^\alpha} V'^\alpha \right) \]

Applying the product rule to the term in the parenthesis, we obtain two separate terms:

\[ \left(\nabla_{\frac{\partial}{\partial t}} V^\mu\right)' = \frac{\partial x'^\mu}{\partial x^\nu} \frac{\partial x^\nu}{\partial x'^\alpha} \frac{dV'^\alpha}{dt} + \frac{\partial x'^\mu}{\partial x^\nu} \frac{d}{dt} \left( \frac{\partial x^\nu}{\partial x'^\alpha} \right) V'^\alpha \]

The first term simplifies using the identity \(\frac{\partial x'^\mu}{\partial x^\nu} \frac{\partial x^\nu}{\partial x'^\alpha} = \delta^\mu_\alpha\):

\[ \frac{\partial x'^\mu}{\partial x^\nu} \frac{\partial x^\nu}{\partial x'^\alpha} \frac{dV'^\alpha}{dt} = \frac{dV'^\mu}{dt} \]

For the second term, we apply the chain rule along the curve: \(\frac{d}{dt} = \frac{dx'^\beta}{dt} \frac{\partial}{\partial x'^\beta}\). This yields:

\[ \left(\nabla_{\frac{\partial}{\partial t}} V^\mu\right)' = \frac{dV'^\mu}{dt} + \left( \frac{\partial x'^\mu}{\partial x^\nu} \frac{\partial^2 x^\nu}{\partial x'^\alpha \partial x'^\beta} \right) \frac{dx'^\beta}{dt} V'^\alpha \tag{3} \]

The first term \(\frac{dV'^\mu}{dt}\) is the naive coordinate rate of change. The second term is a non-tensorial correction that depends on the second derivative of the coordinate transformation. This is precisely the connection term, and the term in parentheses defines the Christoffel symbol.

3. Metric Derivatives and Christoffel Symbols

To express the term in parentheses in terms of the metric tensor \(g_{\alpha\beta}\), we start with the metric transformation law relating our general frame to the flat inertial frame:

\[ g_{\alpha\beta} = \frac{\partial x^\mu}{\partial x'^\alpha} \frac{\partial x^\nu}{\partial x'^\beta} \eta_{\mu\nu} \tag{4} \]

Taking the partial derivative of both sides with respect to a coordinate \(x'^\gamma\), and noting that the derivatives of the flat metric \(\eta_{\mu\nu}\) vanish in the LIF (\(\partial_\gamma \eta_{\mu\nu} = 0\)), we only differentiate the coordinate transformation terms:

\[ g_{\alpha\beta,\gamma} = \frac{\partial^2 x^\mu}{\partial x'^\alpha \partial x'^\gamma} \frac{\partial x^\nu}{\partial x'^\beta} \eta_{\mu\nu} + \frac{\partial x^\mu}{\partial x'^\alpha} \frac{\partial^2 x^\nu}{\partial x'^\beta \partial x'^\gamma} \eta_{\mu\nu} \tag{5} \]

Permuting the indices \(\alpha, \beta, \gamma\), we write the other two partial derivatives:

\[ g_{\gamma\alpha,\beta} = \frac{\partial^2 x^\nu}{\partial x'^\gamma \partial x'^\beta} \frac{\partial x^\mu}{\partial x'^\alpha} \eta_{\mu\nu} + \frac{\partial x^\nu}{\partial x'^\gamma} \frac{\partial^2 x^\mu}{\partial x'^\alpha \partial x'^\beta} \eta_{\mu\nu} \tag{6} \]

\[ g_{\beta\gamma,\alpha} = \frac{\partial^2 x^\nu}{\partial x'^\beta \partial x'^\alpha} \frac{\partial x^\mu}{\partial x'^\gamma} \eta_{\mu\nu} + \frac{\partial x^\mu}{\partial x'^\beta} \frac{\partial^2 x^\nu}{\partial x'^\gamma \partial x'^\alpha} \eta_{\mu\nu} \tag{7} \]

Now, let us form the linear combination \(\frac{1}{2}(g_{\beta\gamma,\alpha} + g_{\gamma\alpha,\beta} - g_{\alpha\beta,\gamma})\). Due to the symmetry of second partial derivatives (e.g., \(\frac{\partial^2 x^\mu}{\partial x'^\alpha \partial x'^\beta} = \frac{\partial^2 x^\mu}{\partial x'^\beta \partial x'^\alpha}\)), four of the six terms on the right-hand side cancel out pairwise, leaving:

\[ \frac{1}{2}(g_{\beta\gamma,\alpha} + g_{\gamma\alpha,\beta} - g_{\alpha\beta,\gamma}) = \eta_{\mu\nu} \frac{\partial^2 x^\nu}{\partial x'^\beta \partial x'^\alpha} \frac{\partial x^\mu}{\partial x'^\gamma} \tag{8} \]

We can rewrite the Minkowski metric \(\eta_{\mu\nu}\) in terms of the general metric \(g_{\Omega\delta}\) using the inverse transformation law: \(\eta_{\mu\nu} = \frac{\partial x'^\Omega}{\partial x^\mu} \frac{\partial x'^\delta}{\partial x^\nu} g_{\Omega\delta}\). Substituting this in yields:

\[ \frac{1}{2}(g_{\beta\gamma,\alpha} + g_{\gamma\alpha,\beta} - g_{\alpha\beta,\gamma}) = \left( \frac{\partial x'^\Omega}{\partial x^\mu} \frac{\partial x'^\delta}{\partial x^\nu} g_{\Omega\delta} \right) \frac{\partial x^\mu}{\partial x'^\gamma} \frac{\partial^2 x^\nu}{\partial x'^\beta \partial x'^\alpha} \]

Using the Kronecker delta identity \(\frac{\partial x'^\Omega}{\partial x^\mu} \frac{\partial x^\mu}{\partial x'^\gamma} = \delta^\Omega_\gamma\), the expression simplifies to:

\[ \frac{1}{2}(g_{\beta\gamma,\alpha} + g_{\gamma\alpha,\beta} - g_{\alpha\beta,\gamma}) = g_{\gamma\delta} \frac{\partial x'^\delta}{\partial x^\nu} \frac{\partial^2 x^\nu}{\partial x'^\beta \partial x'^\alpha} \tag{9} \]

To isolate the coordinate transformation factor, we multiply both sides by the inverse metric \(g^{\mu\gamma}\) and use \(g^{\mu\gamma}g_{\gamma\delta} = \delta^\mu_\delta\):

\[ \frac{\partial x'^\mu}{\partial x^\nu} \frac{\partial^2 x^\nu}{\partial x'^\alpha \partial x'^\beta} = \frac{1}{2} g^{\mu\gamma} (g_{\gamma\alpha,\beta} + g_{\gamma\beta,\alpha} - g_{\alpha\beta,\gamma}) \tag{10} \]

This is exactly the expression for the Christoffel symbol of the second kind, \(\Gamma^\mu_{\alpha\beta}\).

4. Final Result

Comparing Equation (10) back to our transformed derivative in Equation (3), we see that the non-tensorial second-derivative term is precisely the Christoffel symbol \(\Gamma^\mu_{\alpha\beta}\). Therefore, we obtain the full covariant derivative of a vector in our general coordinate system:

                    \[ \left(\nabla_{\frac{\partial}{\partial t}} V^\mu\right)' = \frac{dV'^\mu}{dt} + \Gamma^\mu_{\alpha\beta} \frac{dx'^\beta}{dt} V'^\alpha \tag{11} \]
                

Conclusion: When we differentiate a vector field, the derivative in a locally inertial frame is just the ordinary rate of change. However, when we view this derivative from a general, curved, or accelerating coordinate system, we must apply coordinate transformations. The second-order derivatives of these transformations produce fictitious terms, which are exactly captured by the Christoffel symbols. This confirms that the covariant derivative is the unique, geometrically consistent way to define rates of change on any manifold.

References

Barrett O'Neill, Semi-Riemannian Geometry: With Applications to Relativity, Academic Press, 1983.
Manfredo P. do Carmo, Riemannian Geometry, Birkhäuser, 1992.

Was this post helpful?

Global rating: Loading...

★ ★ ★ ★ ★

Have questions or suggestions? Reach out directly at rprakasht0804@gmail.com