When is a function differentiable? Rules and examples

A graph can be continuous and still have a point where we can’t assign it a slope. Differentiability asks whether the derivative exists as a finite number at the point we’re checking. Let’s connect that definition to corners, vertical tangents, and the places where two pieces of a function meet.

What differentiable at a point means

For an interior point x = c, we use the difference quotient to compare the slopes as a nearby point approaches c. The function is differentiable at c when this limit exists as a finite real number:

f′(c)=limh→0f(c+h)-f(c)h

The slopes from the left and the right must approach the same finite value. That value is the derivative, or the slope of the tangent line, at c.

Differentiability implies continuity

If f is differentiable at c, then f is continuous at c. The function value and the nearby values must fit together before there can be a derivative there, so a hole, a jump, or a vertical asymptote prevents differentiability at that location.

Continuity alone does not give us a derivative. The pieces of a graph can meet while approaching the point with different slopes.

Example: A continuous corner

For f(x) = |x|, the two sides meet at (0, 0), so the function is continuous at 0. To the left, the graph follows y = −x and has slope −1. To the right, it follows y = x and has slope 1.

A continuous corner with two different one-sided slopesThe graph y equals absolute value of x is continuous at the origin. The left slope is negative one, and the right slope is positive one, so the derivative at zero does not exist.y = |x|Left slope: −1Right slope: 1−220xy
The two pieces meet, but their slopes do not match at the corner.

Because those one-sided slopes are different, f′(0) does not exist. The function is differentiable on (−∞, 0) and on (0, ∞), but not at 0.

Vertical tangents, cusps, and discontinuities

For y=x13, the secant slopes at the origin become unbounded. The graph has a vertical tangent there, but there is no finite derivative at 0. A cusp can also have unbounded one-sided slopes, as happens for y=|x|23 at 0.

These are common graph features to investigate. The deciding test is still the finite limit in the derivative definition; continuity and a graph that looks smooth are not a substitute for that limit.

Example: Make two polynomial pieces differentiable

Let’s find a and b so this piecewise function is differentiable where the two pieces meet:

f(x)={2x2-ax+bx≤13x3-x2-6x>1

The pieces change at x = 1. First, their values must meet there. The left piece includes x = 1, so matching it with the right-hand limit also makes the common limit equal f(1).

2-a+b=3-1-6
-a+b=-6

Next, the one-sided derivatives must match. Since both pieces are polynomials, we can evaluate the derivative of each piece at 1:

4-a=9-2=7
a=-3

Substitute a = −3 into the continuity equation to find b.

-(-3)+b=-6⟹b=-9

With these constants, both pieces have value −4 at 1 and both one-sided derivatives equal 7. The function is therefore differentiable at the join.

Differentiability on an interval and at endpoints

A function is differentiable on an open interval when it has a finite derivative at every point in that interval. If a problem asks about an endpoint of the domain, there is only one available side, so we state explicitly when we’re using a one-sided derivative. This distinguishes endpoint behavior from the usual two-sided derivative at an interior point.

Video music: “Wings” by Nicolai Heidlas.