The Chain Rule
When one function sits inside another, rates of change multiply. Differentiate the outer function, leaving the inside alone, then multiply by the derivative of the inside.
What the symbols mean
- Outer derivative
Differentiate the outside function, leaving the inside alone.
- Inner derivative
How fast the inside changes.
What the symbols mean
- Inside variable
u = g(x). Rates multiply along the chain.
Explanation
The chain rule is simply how you differentiate a function that's inside another function, like or . The name comes from thinking of it as a chain of links: feeds into the inside function, and the inside feeds into the outside. All it really says is that rates multiply along the chain. The formula looks crowded, but it's two pieces: (how fast the outside changes, with the inside left untouched) times (how fast the inside changes). That's the whole rule, and it's the one you'll use more than any other.
Rates multiply. Here's why it should be a product. Suppose changes 3 times as fast as , and changes 2 times as fast as . Nudge by a little, and moves twice as much. That makes move three times as much as did, so moves times as much as . In Leibniz notation it almost looks like fractions canceling: . (They aren't really fractions, but it's a helpful way to remember it.)
So far, every rule we've had needed the function to be a plain power, exponential or trig function of . The chain rule is what lets all of them work when is replaced by something more complicated.
How to apply it. Name the outside and the inside. Differentiate the outside as if the inside were a single variable, then multiply by the inside's derivative.
A common mistake. It's tempting to differentiate the outside and stop, because that's what the old rules trained you to do. So comes out as . It isn't: the inside changes 5 times as fast as , so the answer is . Forgetting the inner derivative is the single most common calculus mistake there is.
Patterns worth memorizing:
| Function | Derivative |
|---|---|
For a chain of three or more, keep peeling layers from the outside in, one rate per layer: .
See it
Worked examples
Differentiate .
- 1
The outside is , with derivative . The inside is , with . Multiply them, putting the inside back in for .
Differentiate .
- 1
The outside is , which is its own derivative, so it stays . The inside is , with derivative . Multiply the two.
Practice
Type your answer and check it. Solve them all to complete the lesson.
Differentiate .
Differentiate .
Differentiate .
Differentiate .
Differentiate .
Differentiate .
A balloon's radius grows as cm, and its volume is . How fast is the volume growing at s?
The hardest one here. Get it, and you've got this lesson.
Differentiate .
Where it's used
Backpropagation, the algorithm that trains neural networks, is the chain rule applied through millions of nested functions. Related rates (two lessons on) is the chain rule with time as the variable.