Course contentsShow
← Mathematics
Lesson 1861 of 3,12139. Multivariable Functions and Partial DerivativesFree lesson

General Chain Rule Formula and Notation

Master the general formulation of the multivariable chain rule using summation notation and Jacobian matrix concepts.

General Chain Rule Formula and Notation

You'll learn how to express the multivariable chain rule in its most general form using summation notation and the Jacobian matrix framework.

The Big Picture

When a multivariable function depends on intermediate variables that themselves depend on other variables, the chain rule tells us how to compute derivatives with respect to the final independent variables. The general formula unifies all the special cases you've seen (paths, implicit differentiation, change of variables) into one systematic notation.

The General Formula

If z = f(u₁, u₂, ..., uₘ) and each uᵢ = uᵢ(t₁, t₂, ..., tₙ), then:

∂z∂tj=∑i=1m∂f∂ui⋅∂ui∂tj\frac{\partial z}{\partial t_j} = \sum_{i=1}^{m} \frac{\partial f}{\partial u_i} \cdot \frac{\partial u_i}{\partial t_j}

This says: to find how z changes with respect to any independent variable tⱼ, sum over all intermediate variables, multiplying the rate of change of f with respect to each uᵢ by the rate of change of uᵢ with respect to tⱼ.

Jacobian Matrix Perspective

We can organize this as a matrix product. The derivative of z with respect to t = (t₁, ..., tₙ) is:

∇tz=∇uf⋅Ju(t)\nabla_{\mathbf{t}} z = \nabla_{\mathbf{u}} f \cdot J_{\mathbf{u}}(\mathbf{t})

where J is the Jacobian matrix with entries ∂uᵢ/∂tⱼ. This compactly captures all partial derivatives at once.

Why This Matters

This notation reveals the structure: tree diagrams become matrix multiplications. Every path from z to tⱼ through intermediate variables contributes one term to the sum. It's like tracking all possible routes through a network—each route's contribution multiplies the rates along that path.

Key Takeaway: The general chain rule expresses ∂z/∂tⱼ as a sum over all intermediate variables, or equivalently as a Jacobian matrix product, unifying all multivariable differentiation patterns into one formula.