Computing a derivative at , then again at , then at gets old fast. It is better to do the computation once with a general and get a formula that gives the slope at every point. That formula is a function in its own right, and studying where it exists tells you exactly where a graph is smooth.
Key terms
The derivative function
Definition For a function , the derivative of is the function given by
at every for which is defined on an open interval around and this limit exists as a finite number. The domain of is the set of all such , which can be smaller than the domain of .
In plain words Plug in a general point instead of a specific one, and the derivative becomes a slope formula for the whole graph.
Example For ,
so for every real . For the same method gives , but only for . The domain of is while the domain of is .
Leibniz and prime notation
Definition If , each of the following denotes the derivative.
The value at a specific point is written or . The symbol is a single symbol for a limit, not a fraction of two numbers, although it often behaves like one in calculations.
In plain words Prime notation is short. Leibniz notation shows which variable is changing, which helps when several variables are in play.
Example If then , and . All three say the same thing, and the last one is the slope at the point .
The differentiation operator
Definition The differentiation operator , also written , takes a function and returns the function , defined wherever is differentiable. It acts on whole functions, not on single numbers.
In plain words Think of as a machine. A function goes in, its derivative function comes out.
Example Applying the operator to gives . Applying it again gives , which can be checked by the same limit,
Note that is the derivative of a constant, not .
One-sided derivatives
Definition Suppose is defined on for some . The right-hand derivative of at is
Suppose is defined on for some . The left-hand derivative of at is
Each one exists only when its limit exists as a finite number. For defined on an open interval around , the derivative exists exactly when both one-sided derivatives exist and are equal, and then is their common value.
In plain words Measure the slope using only points to the right, then only points to the left. A true derivative needs both answers to agree.
Example For at the quotient is , which is for and for . So and . They differ, so is not differentiable at .
Differentiability on an interval
Definition A function is differentiable on an open interval , which may be unbounded, if it is differentiable at every point of it. It is differentiable on a closed interval if it is differentiable on and the one-sided derivatives and exist. Half-open intervals are handled the same way at the included endpoint.
In plain words Inside the interval use ordinary derivatives. At an included endpoint only the slope from inside is required.
Example The function is differentiable on , with and . The function is differentiable on but not on , because at the quotient tends to as , so does not exist.
Differentiability implies continuity
Definition If is differentiable at , then is continuous at . Proof. For ,
so . The converse is false, since a function can be continuous at a point without being differentiable there.
In plain words A graph that has a tangent line at a point cannot have a break there. But an unbroken graph can still fail to have a tangent.
Example The floor function jumps at , so it is not continuous there, and therefore it cannot be differentiable at . In the other direction, is continuous at but not differentiable there.
Corners
Definition Let be defined on an open interval around . Then has a corner at if the one-sided derivatives and both exist as finite numbers and are different. Then is continuous at , because each one-sided derivative forces continuity from that side, but does not exist. Some books also call a point a corner when the slope is finite on one side and infinite on the other.
In plain words The graph arrives along one straight direction and leaves along another, making a sharp bend.
Example Let at , where . For small we have , and for small we have . So the quotients are
So , , and the graph has a corner at .
Cusps
Definition Let be defined on an open interval around and continuous at . Then has a cusp at if the difference quotient tends to from one side of and to from the other side. That is, one of the limits
is and the other is .
In plain words Both sides of the graph become vertical as they meet, but one comes down and the other goes up, forming a sharp point. Continuity matters here, since without it a jump could also make the quotients blow up.
Example For at , the quotient is . As this tends to , and as it tends to , since the cube root of a small negative number is a small negative number. The graph has a cusp at the origin.
Vertical tangents
Definition Let be defined on an open interval around and continuous at . Then has a vertical tangent at if the difference quotient tends to from both sides of , or to from both sides. The tangent line is then the vertical line , and does not exist.
In plain words The curve passes straight up or straight down for an instant. A vertical line has no slope, so there is no derivative.
Example For at the quotient is
which is positive and tends to as from either side. So the graph has the vertical tangent at the point .
Check yourself
- Use the limit definition to find for .
Show the answer
The difference quotient simplifies to
so for every .
- Find and for . Is differentiable at ?
Show the answer
For , , slope . For , , slope . So and . They differ, so is not differentiable at . It has a corner there.
- Classify the point for and for .
Show the answer
For the quotient is from both sides, a vertical tangent. For the quotient is , so and is differentiable there.