From: 3blue1brown
Implicit differentiation is a technique used in calculus to find the slope of a tangent line to a curve or the rate of change of variables in an equation where one variable is not explicitly defined as a function of the other [01:31:00]. This method is particularly useful for “implicit curves,” which are sets of points (x, y) that satisfy a property written in terms of two variables, x and y [01:52:04].
Finding the Slope of a Tangent Line to a Circle
Consider a circle with radius 5 centered at the origin, defined by the equation x² + y² = 5² [00:00:22]. To find the slope of a tangent line to this circle, for instance at the point (3,4) [00:00:44], a direct application of simple derivatives is not possible because the curve is not the graph of a single function [01:31:00]. X is not an input, and y is not an output; they are interdependent values [01:44:26].
The key idea for finding tangent line slopes is to “zoom in” until the curve looks like its tangent line, then consider a tiny step along the curve [01:07:11]. The y-component of this step is dy
, and the x-component is dx
, making the desired slope dy/dx
[01:17:21].
To apply implicit differentiation:
- Take the derivative of both sides of the equation x² + y² = 5² [02:12:12].
- For x², the derivative is
2x * dx
[02:16:32]. - For y², the derivative is
2y * dy
[02:19:50]. - The derivative of the constant 5² is 0 [02:24:25].
- This yields the equation:
2x dx + 2y dy = 0
[02:46:17].
- For x², the derivative is
- Rearrange the equation to find an expression for
dy/dx
[02:46:17].2y dy = -2x dx
dy / dx = -x / y
[02:51:24].
At the point (x, y) = (3, 4), the slope dy/dx
is -3/4 [02:56:56].
Connection to Related Rates Problems
Implicit differentiation is conceptually linked to “related rates” problems [03:20:00].
Example: The Ladder Problem
Imagine a 5-meter ladder leaning against a wall [03:26:34]. If the top of the ladder is 4 meters above the ground and is slipping down at 1 meter per second, the bottom is initially 3 meters from the wall [03:30:19]. The question is, at what rate is the bottom of the ladder moving away from the wall at that moment [03:46:42]?
- Define variables: Let
y(t)
be the distance from the top of the ladder to the ground, andx(t)
be the distance from the bottom of the ladder to the wall [04:21:49]. - The relationship is given by the Pythagorean theorem:
x(t)² + y(t)² = 5²
[04:34:01]. This equation holds true at all times [04:43:40]. - Take the derivative of both sides with respect to time (
t
) [05:36:58].- The left side
x(t)² + y(t)²
is a function of time [05:17:09]. - Using the chain rule:
d/dt (x(t)²) = 2 * x(t) * (dx/dt)
[06:08:42]d/dt (y(t)²) = 2 * y(t) * (dy/dt)
[06:27:06]
- The right side,
5²
(a constant), has a derivative of 0 [05:53:57]. - This results in:
2x (dx/dt) + 2y (dy/dt) = 0
[06:35:54].
- The left side
- Substitute known values at
t=0
:y(t) = 4m
,x(t) = 3m
, anddy/dt = -1 m/s
(negative because it’s dropping) [06:45:51].2(3) (dx/dt) + 2(4) (-1) = 0
6 (dx/dt) - 8 = 0
dx/dt = 8/6 = 4/3 meters per second
[07:04:15].
In the ladder problem, taking the derivative has a clear meaning: it’s the rate at which the expression changes as time changes [07:32:04]. For the circle, the “tiny nudges dx
and dy
” are not tied to an external variable like time [07:52:16], which initially feels strange [07:43:08].
Interpretation of Implicit Differentiation
To understand the meaning of taking a derivative of an expression with multiple variables:
- Let
s = x² + y²
[08:03:07]. Thiss
is a function of two variables,x
andy
[08:08:11]. For points on the circle,s
equals 25 [08:16:38]. - Taking a derivative of this expression means considering a tiny change
dx
tox
and a tiny changedy
toy
[08:34:03]. - The change in
s
, denotedds
, is approximately2x dx + 2y dy
[09:21:40]. This formula tells you how much the value ofx² + y²
changes based on your starting point (x,y) and the tiny step (dx, dy) [09:41:43]. - When you restrict yourself to steps along the circle, you are ensuring that the value of
s
does not change; it remains 25 [10:07:00]. - Therefore,
ds
must be 0 [10:17:34]. - Setting the expression
2x dx + 2y dy
equal to 0 is the condition for a tiny step to stay on the tangent line of the circle [10:20:23]. For tiny enough steps, the tangent line and the circle are essentially the same [10:40:02].
Generalizing Implicit Differentiation
Implicit differentiation can be applied to any equation relating x
and y
.
Example: sin(x)y² = x
Consider the equation sin(x)y² = x
, which represents a set of u-shaped curves [10:53:23].
- Take the derivative of each side with respect to
x
(implicitly) [11:23:42].- For the left side,
sin(x)y²
, apply the product rule: “left d right plus right d left” [11:32:06].sin(x) * (change to y²)
, which issin(x) * 2y dy
[11:39:10].y² * (change to sin(x))
, which isy² * cos(x) dx
[11:44:03].
- For the right side,
x
, the change is simplydx
[11:52:16].
- For the left side,
- Set the changes of both sides equal to each other to ensure the step stays on the curve [11:59:04].
sin(x) (2y dy) + y² cos(x) dx = dx
[11:59:04].
- From here, you can algebraically rearrange to find
dy/dx
[12:26:17].
Deriving New Derivative Formulas
Implicit differentiation can also be used to derive new derivative formulas.
Example: Derivative of Natural Logarithm ln(x)
To find the derivative of ln(x)
:
- Let
y = ln(x)
[12:50:00]. This implicitly defines the curve. - Rearrange the equation using the definition of natural logarithm:
e^y = x
[13:16:16]. - Take the derivative of both sides, considering how a tiny step (
dx
,dy
) changes the value of each side [13:31:00].d/dx (e^y)
using the chain rule ise^y dy
[13:44:16].d/dx (x)
isdx
[13:50:20].
- Equate the changes to ensure the step stays on the curve:
e^y dy = dx
[13:50:20]. - Rearrange to solve for
dy/dx
:dy/dx = 1 / e^y
[13:57:38]. - Since
e^y
is equal tox
when on the curve, substitutex
fore^y
:dy/dx = 1 / x
[14:06:17].
This demonstrates that the derivative of ln(x)
is 1/x
[14:24:20].
Implicit differentiation is a fundamental concept in multivariable calculus that helps understand how functions with multiple inputs change when those inputs are tweaked [14:32:00].