How To Know If Function Is Differentiable

Does your function have a tangent line that actually touches it cleanly? Or are you just assuming it does because it looks smooth?

Here's what most people miss: continuity alone doesn't cut it. That's why a function can be continuous everywhere but still have corners, cusps, or vertical tangents that kill differentiability. So how do you actually know if a function is differentiable?

What Does It Mean For a Function to Be Differentiable?

A function is differentiable at a point if its derivative exists there. That sounds circular, but here's what it actually means in practice: you can draw a unique tangent line at every point on the curve, and that tangent line has a well-defined slope.

More formally, the limit that defines the derivative must exist. But this means the left-hand and right-hand limits of the difference quotient must agree and be finite. No jumping, no blowing up to infinity, no disagreement about what the slope should be.

Think of it this way: if you're zooming in on the graph at a point, a differentiable function will eventually look like a straight line. Non-differentiable points are where this breaks down—sharp corners, vertical drops, or oscillating wiggles that never settle down.

Differentiability vs Continuity

Here's the crucial relationship: every differentiable function is continuous, but not every continuous function is differentiable. This trips people up all the time.

The absolute value function f(x) = |x| is the classic example. But at x = 0, there's a sharp corner. Because of that, it's continuous everywhere—you can draw it without lifting your pen. No unique tangent line exists there, so it's not differentiable at that point.

This is why checking continuity is usually step one, but it's not the finish line.

Why Knowing Differentiability Matters

Differentiability isn't just mathematical housekeeping—it's the gateway to powerful tools. When a function is differentiable, you can:

Use linear approximation to estimate values nearby
Apply the chain rule, product rule, and all of calculus' machinery
Solve optimization problems with confidence
Model real-world rates of change accurately

In physics, engineering, economics—anywhere rates matter—you need differentiability to make meaningful calculations. A position function that isn't differentiable means your velocity calculation breaks down. A cost function with kinks might hide important behavioral thresholds.

How to Actually Determine If a Function Is Differentiable

Method 1: Compute the Derivative Directly

This is the straightforward approach. Find f'(x) using whatever differentiation rules apply, then check where it exists.

For polynomials, trigonometric functions, exponentials—these are differentiable everywhere on their domains. Their derivatives exist and are continuous.

For rational functions, watch out for division by zero. f(x) = 1/x is differentiable everywhere except x = 0.

Piecewise functions require extra care. You need to check differentiability at the boundary points where the formula changes. Worth knowing.

Method 2: Use the Definition of the Derivative

Sometimes computing the derivative is messy or impossible algebraically. That's when you go back to the limit definition:

f'(a) = lim[h→0] (f(a+h) - f(a))/h

The key is checking whether this limit exists. Both the left-hand limit (h→0⁻) and right-hand limit (h→0⁺) must exist, be finite, and equal each other.

Method 3: Check for Common Problem Patterns

Certain red flags almost always signal non-differentiability:

Sharp corners: Absolute values, piecewise linear functions with different slopes on each side

Vertical tangents: Functions like f(x) = ∛x at x = 0—the derivative blows up to infinity

Cusps: Points where the function is continuous but the derivative approaches +∞ from one side and -∞ from the other

Oscillating discontinuities: Functions that wiggle faster and faster near a point, like f(x) = x sin(1/x) near x = 0

Common Mistakes People Make

Assuming Continuity Equals Differentiability

We're talking about the big one. I see students—and frankly, many published explanations—treat continuity as sufficient for differentiability. It's not.

The function f(x) = x² sin(1/x) for x ≠ 0 and f(0) = 0 is continuous at x = 0 but not differentiable there. The oscillations get too wild as you approach zero, so the derivative limit doesn't settle on a single value.

Ignoring Domain Restrictions

A function might be differentiable on its domain, but that domain could exclude important points. f(x) = √x is differentiable on (0, ∞), but not at x = 0—the derivative involves 1/√x, which blows up.

Some sources say "√x is not differentiable at 0"—technically true, others say "it's differentiable on its domain"—also true. The context matters.

Forgetting One-Sided Derivatives

With piecewise functions, you need to check that the left-hand derivative matches the right-hand derivative at transition points.

Consider f(x) = {x² for x ≤ 0, x for x > 0}. At x = 0:

Left-hand derivative: d/dx[x²] = 2x, so at x = 0, it's 0
Right-hand derivative: d/dx[x] = 1

Since 0 ≠ 1, the function isn't differentiable at x = 0.

Overlooking Implicit Assumptions

When a problem says "show f is differentiable," it usually means "show f' exists and is finite." But sometimes the question is really about showing f' is continuous, which is a stronger condition called "continuously differentiable."

Practical Tips That Actually Work

Start with the Graph

If you can see the function, look for obvious trouble spots. Corners, cusps, vertical tangents, discontinuities—all of these kill differentiability. A quick sketch often reveals where you need to focus your analysis.

Check Piecewise Boundaries First

For functions defined in pieces, the boundaries are where problems most often hide. Always compute both one-sided derivatives and verify they match.

Use Algebraic Simplification

When applying the limit definition, try to simplify the difference quotient before taking the limit. Rationalizing numerators, factoring, canceling terms—these tricks often reveal whether the limit exists.

make use of Known Results

Build a mental catalog of standard functions and their differentiability:

Polynomials: differentiable everywhere
Trig functions: differentiable on their domains
Exponentials and logs: differentiable on their domains
Absolute value: not differentiable at the zero point

Test with Specific Values

When in doubt, plug in numbers. Compute the difference quotient for small values of h approaching from both sides. If the quotients are trending toward different values, the derivative doesn't exist.

When Things Get Tricky

Some functions push the boundaries of what "differentiable" even means.

Functions Continuous Everywhere, Differentiable Nowhere

Yes, these exist. In practice, it's a fractal-like object that's always wiggling too much to have a tangent line. The Weierstrass function is continuous everywhere but differentiable nowhere. Fortunately, these rarely appear in applied contexts, but they're important to know about.

Continue exploring with our guides on do diagonals of a parallelogram bisect each other and find the area bounded by the curve.

Differentiability at Isolated Points

A function can be differentiable at a single point but nowhere else nearby. That said, consider f(x) = x² for rational x, 0 for irrational x. At x = 0, the derivative exists (it's 0), but at any other point, the function isn't even continuous, so it can't be differentiable.

Higher-Order Differentiability

Once you know a function is differentiable, you can ask if its derivative is also differentiable. Here's the thing — a function that's twice differentiable has a continuous second derivative. This gives you second derivatives, third derivatives, and so on. These concepts matter in Taylor series, differential equations, and optimization theory.

FAQ

Can a function be differentiable at just one point?

Yes. The standard example is f(x) = x² for x rational, 0 for x irrational. At x = 0, you can show the derivative exists and equals 0, but the function isn't even continuous anywhere else.

If a function is differentiable, is its derivative always continuous?

No. Now, there are functions whose derivatives exist everywhere but aren't continuous at certain points. These are called "differentiable but not continuously differentiable." The derivative exists but has jump discontinuities.

How do I check differentiability for parametric curves?

For a parametric curve defined by x(t) and y(t), the curve is differentiable at t = t₀ if

Differentiability for Parametric Curves

When a curve is described by two functions of a parameter—(x = x(t)) and (y = y(t))—the notion of differentiability shifts from a single‑variable function to a vector‑valued function (\mathbf{r}(t)=\langle x(t),,y(t)\rangle).

A point (t=t_{0}) on the curve is said to be differentiable if the limit

[ \lim_{h\to 0}\frac{\mathbf{r}(t_{0}+h)-\mathbf{r}(t_{0})}{h} ]

exists. Because the numerator is a vector, the limit must exist component‑wise:

[ \lim_{h\to 0}\frac{x(t_{0}+h)-x(t_{0})}{h}=x'(t_{0}),\qquad \lim_{h\to 0}\frac{y(t_{0}+h)-y(t_{0})}{h}=y'(t_{0}) ]

Both one‑sided limits must agree, and the resulting vector (\langle x'(t_{0}),,y'(t_{0})\rangle) is the tangent vector at that point.

When Does the Limit Fail?

Cusp or corner – If (x'(t_{0})) and (y'(t_{0})) exist but are not proportional (i.e., the direction of the tangent vector changes abruptly), the curve has a corner.
Vertical tangent – If (x'(t_{0})=0) while (y'(t_{0})\neq 0), the limit vector points straight up or down, giving a vertical tangent line.
Higher‑order degeneracy – If both derivatives vanish, higher‑order terms must be examined. Take this case: the cycloid (x(t)=\sin t,;y(t)=1-\cos t) has a cusp at (t=0) because the first non‑zero derivative appears at second order.

Implicit Differentiation and Differentiability

Often a curve is given implicitly by an equation (F(x,y)=0). If (F) is differentiable and (\partial F/\partial y\neq 0) at a point ((x_{0},y_{0})), the Implicit Function Theorem guarantees a locally unique function (y=g(x)) that is differentiable there, and its derivative is

[ g'(x)=\frac{-\partial F/\partial x}{\partial F/\partial y}. ]

If the denominator vanishes, the curve may still be differentiable in a parametric sense, but the explicit formula for (dy/dx) fails; instead, one must resort to the parametric condition above.

Extending the Idea to Several Variables

Differentiability is not confined to curves in the plane. In multivariable calculus, a function (F:\mathbb{R}^{n}\to\mathbb{R}^{m}) is differentiable at a point (\mathbf{a}) if there exists a linear map (L:\mathbb{R}^{n}\to\mathbb{R}^{m}) such that

[ \lim_{\mathbf{h}\to\mathbf{0}}\frac{|F(\mathbf{a}+\mathbf{h})-F(\mathbf{a})-L(\mathbf{h})|}{|\mathbf{h}|}=0. ]

When (m=1) this reduces to the familiar gradient notion; when (m>1) the linear map is represented by the Jacobian matrix. The existence of such a linear approximation is equivalent to all partial derivatives existing and satisfying a certain continuity condition that prevents “corner‑like” behavior in higher dimensions.

Practical Takeaways

Look for smoothness – If a function can be expressed using elementary operations (polynomials, exponentials, trigonometric functions, compositions) on domains where those operations are defined, it is almost always differentiable there.
Spot the trouble spots – Points where the definition changes (absolute value, piecewise formulas), where denominators vanish, or where radicals are taken of negative numbers are prime candidates for non‑differentiability.
Use limits wisely – Compute one‑sided limits of the difference quotient; if they disagree, the derivative does not exist.
Remember parametric nuance – For curves defined parametrically, differentiability hinges on the existence of a common limit for both component derivatives, and the resulting tangent vector may reveal hidden geometry such as cusps or vertical tangents.
Higher‑order considerations – When a function is differentiable, you may ask whether its derivative is itself differentiable. Repeating this process yields higher‑order derivatives, which are essential for series expansions, optimization, and the study of curvature in geometry.

Conclusion

Understanding differentiability is tantamount to recognizing where a function behaves “smoothly” enough to admit a linear approximation—a tangent line, a tangent plane, or a tangent vector in more abstract settings. By systematically checking continuity, employing algebraic tricks to simplify difference quotients, leveraging known differentiable building blocks, and probing special cases such as

and probing special cases such as cusps, vertical tangents, or sudden changes in direction—situations where the usual algebraic criteria for differentiability break down but the geometric intuition of a well‑behaved tangent still guides the analysis.

A Few Final Observations

Cusp vs. Corner – At a cusp the tangent direction exists but the speed along the curve changes sign; at a corner the direction itself jumps. Both are detectable via the one‑sided limits of the derivative or, in parametric form, by comparing the limits of the ratio (dy/dx).
Vertical Tangents – When (dx/dt=0) while (dy/dt\neq0), the tangent line is vertical. The derivative (dy/dx) is formally infinite, yet the parametric definition still yields a perfectly valid tangent vector.
Implicit Functions – When an equation defines (y) implicitly in terms of (x), the derivative can be obtained by differentiating both sides and solving for (dy/dx). The implicit function theorem guarantees differentiability provided the partial derivative with respect to (y) does not vanish.

These nuanced scenarios reinforce that differentiability is not merely an algebraic convenience; it is a geometric property that reflects the local linearity of a function or curve. When the linear approximation exists, calculus can proceed with confidence—whether we are computing rates of change, optimizing functions, or linearizing nonlinear systems for numerical analysis.

Final Word

Differentiability is the bridge between the discrete world of algebraic expressions and the continuous realm of geometry. Here's the thing — in practice, the quest for differentiability often boils down to checking continuity, simplifying limit expressions, and recognizing the geometric signatures of non‑smooth behavior. By demanding a linear approximation at every point, we enforce a form of smoothness that underpins the entire edifice of differential calculus. Whether one is tracing a parametric curve, solving an optimization problem in several variables, or simply verifying the behavior of a piecewise function, the principles outlined above provide a reliable toolkit. Armed with these strategies, one can confidently handle the landscape of functions, confident that wherever the derivative exists, a well‑defined tangent—be it a line, a plane, or a vector—awaits.