The Function That Throws Away Everything But The Sign
Most functions care about how big a number is. The signum function ignores size entirely and reports one thing: is the input negative, zero, or positive? That deliberate forgetfulness is exactly what makes it useful.
Whenever a rule needs "which direction" without "how far" - a controller that must push left or right, a sorting comparison, an absolute value rebuilt from a sign - the signum function is the tool that strips a number down to its direction.
What Is The Signum Function?
The signum function (from the Latin signum, "sign") is the function $\operatorname{sgn} : \mathbb{R} \to {-1, 0, 1}$ that returns the sign of its input. Its piecewise definition is:
$$\operatorname{sgn}(x) = \begin{cases} -1 & \text{if } x < 0, \ 0 & \text{if } x = 0, \ 1 & \text{if } x > 0. \end{cases}$$
So $\operatorname{sgn}(-8) = -1$, $\operatorname{sgn}(0) = 0$, and $\operatorname{sgn}(3.7) = 1$. The exact input value is discarded; only its sign survives. A closely related identity ties it to the absolute value: for every nonzero $x$,
$$\operatorname{sgn}(x) = \frac{x}{|x|}, \qquad x \neq 0.$$
This says the signum of $x$ is $x$ divided by its own magnitude, which is why it always lands on $-1$ or $+1$ away from the origin. (At $x = 0$ the fraction is undefined, so the value $0$ is assigned separately.)
What Are The Domain And Range Of The Signum Function?
The signum function accepts every real number, so its domain is all of $\mathbb{R}$. But it only ever outputs one of three values, so its range is the finite set
$$\text{Range} = {-1, 0, 1}.$$
That mismatch - an infinite domain squeezed onto three outputs - is the source of every structural property below. Because many different inputs (say $2$, $50$, and $10^6$) all map to the same output $1$, the signum function cannot be one-to-one.
What Does The Graph Of The Signum Function Look Like?
The graph has three disconnected pieces:
A horizontal ray at height $y = -1$ for every $x < 0$.
A single point at the origin, $(0, 0)$, where $\operatorname{sgn}(0) = 0$.
A horizontal ray at height $y = +1$ for every $x > 0$.
At $x = 0$ the value leaps from $-1$ to $+1$, a jump discontinuity of size $2$. The function is flat on each side (its slope is $0$ everywhere it is defined off the origin), so nothing "curves"; it simply sits at one level, jumps, and sits at another. This shape puts the signum function in the same family as the greatest integer function, another step-style function built from flat pieces and jumps.
What Are The Properties Of The Signum Function?
The signum function's small range gives it a tidy list of properties.
It is an odd function. $\operatorname{sgn}(-x) = -\operatorname{sgn}(x)$ for all $x$, so its graph has rotational symmetry about the origin. This places it among the even and odd functions.
It is neither one-to-one nor onto (as a map to $\mathbb{R}$). Many inputs share an output, and it hits only three real values.
It is multiplicative: $\operatorname{sgn}(xy) = \operatorname{sgn}(x),\operatorname{sgn}(y)$. The sign of a product is the product of the signs.
It rebuilds absolute value: $|x| = x \cdot \operatorname{sgn}(x)$, and conversely $\operatorname{sgn}(x) = \frac{x}{|x|}$ for $x \neq 0$.
It is bounded: $-1 \le \operatorname{sgn}(x) \le 1$ for every real $x$.
It is discontinuous only at $x = 0$, and constant (hence continuous) everywhere else.
What Is The Derivative Of The Signum Function?
Away from the origin the signum function is flat, so its derivative is $0$ for every $x \neq 0$:
$$\frac{d}{dx}\operatorname{sgn}(x) = 0, \qquad x \neq 0.$$
At $x = 0$ the ordinary derivative does not exist, because a function must be continuous at a point to be differentiable there, and the signum function jumps at $0$. (In more advanced work the "derivative" at the origin is described by the Dirac delta, but at school level the honest statement is: slope $0$ off the origin, undefined at the origin.) This is a gap most competitor pages skip, and it is worth stating plainly rather than leaving a student to assume the derivative is $0$ everywhere.
Examples Of Signum Function
The set runs from a direct evaluation, through the mistake students make at $x = 0$, up to solving an inequality and using the multiplicative property.
Example 1
Evaluate $\operatorname{sgn}(-15)$, $\operatorname{sgn}(0)$, and $\operatorname{sgn}(4.2)$.
Read the sign of each input against the definition.
$$\operatorname{sgn}(-15) = -1$$ $$\operatorname{sgn}(0) = 0$$ $$\operatorname{sgn}(4.2) = 1$$
Final answer: $-1$, $0$, and $1$. Magnitude is irrelevant; $-15$ and $-0.0001$ both give $-1$.
Example 2
Find $\operatorname{sgn}(x)$ when $x = 3 - 3$.
Wrong path. A student sees the positive-looking numbers $3$ and $3$ and answers $\operatorname{sgn}(x) = 1$, reasoning "the inputs are positive, so the sign is positive." But the signum function reads the value of $x$, not the numbers used to build it. First simplify.
Correct. Evaluate the input before applying the function.
$$x = 3 - 3 = 0$$ $$\operatorname{sgn}(0) = 0$$
Final answer: $0$. Always compute the argument first; $\operatorname{sgn}$ then reads that single resulting value.
Example 3
Simplify $\operatorname{sgn}(x) \cdot |x|$ for a nonzero $x$.
Use the identity $|x| = x \cdot \operatorname{sgn}(x)$ and the fact that $\operatorname{sgn}(x)^2 = 1$ for $x \neq 0$.
$$\operatorname{sgn}(x) \cdot |x| = \operatorname{sgn}(x)\cdot\big(x\cdot\operatorname{sgn}(x)\big)$$ $$= x \cdot \operatorname{sgn}(x)^2$$ $$= x \cdot 1 = x$$
Final answer: $x$. The signum and the absolute value are inverse "sign-strippers": multiply them and the original number returns.
Example 4
Use the multiplicative property to find $\operatorname{sgn}(-6 \cdot 5)$.
The sign of a product is the product of the signs.
$$\operatorname{sgn}(-6) = -1$$ $$\operatorname{sgn}(5) = 1$$ $$\operatorname{sgn}(-6 \cdot 5) = (-1)(1) = -1$$
Final answer: $-1$. You never have to multiply $-6$ by $5$; the signs alone settle it.
Example 5
Solve $\operatorname{sgn}(x - 4) = 1$.
The output is $1$ exactly when the input is strictly positive.
$$\operatorname{sgn}(x - 4) = 1 \iff x - 4 > 0$$ $$x > 4$$
Final answer: $x > 4$. Reading the definition backwards turns a signum equation into a simple inequality.
Example 6
Find $\operatorname{sgn}(x^2 + 1)$ for every real $x$.
Check the sign of the input across all real $x$.
$$x^2 \ge 0 \text{ for all real } x$$ $$x^2 + 1 \ge 1 > 0$$ $$\operatorname{sgn}(x^2 + 1) = 1$$
Final answer: $1$, for every real $x$. Because $x^2 + 1$ can never be zero or negative, the signum is constantly $+1$.
Where The Signum Function Earns Its Keep
"How does a machine know which way to push before it knows how hard?"
Control systems answer that question with a signum. A bang-bang controller - the logic behind a simple thermostat or a satellite's attitude thrusters - does not compute a smooth response; it looks at the sign of the error and pushes full force in the opposite direction. If the error is positive, fire one way; if negative, fire the other. That is $\operatorname{sgn}(\text{error})$ doing the deciding. The function's whole value is that it separates direction from magnitude, which is often exactly the split an engineer needs.
The signum function also shows up quietly across mathematics: it appears in the formula for the derivative of $|x|$ (which is $\operatorname{sgn}(x)$ away from zero), in defining orientation and permutations by their "sign," and anywhere a formula must branch on positive-versus-negative without caring about the exact number. It belongs to the broader family of types of functions that students meet in higher algebra, and it is one of the cleanest examples of a piecewise, discontinuous rule.
Common Mistakes
Mistake 1: Forgetting That sgn(0) Is 0, Not 1 Or −1
Where it slips in: any input that simplifies to zero.
Don't do this: assign $\operatorname{sgn}(0) = 1$ (or $-1$) because "it has to be one of them." Zero has no sign, and the definition gives it the value $0$.
The correct way: treat $x = 0$ as its own case with output $0$. The memoriser who learned "signum gives plus or minus one" freezes here — the fix is to picture the graph, where the origin is a lone dot at height $0$, distinct from both rays.
Mistake 2: Applying sgn Before Simplifying The Input
Where it slips in: expressions like $\operatorname{sgn}(3 - 3)$ or $\operatorname{sgn}(2 - 5)$.
Don't do this: judge the sign from the numbers you see rather than the value they produce. $\operatorname{sgn}(2 - 5)$ is not "mixed" or "positive"; the input is $-3$.
The correct way: compute the argument to a single number first, then read its sign. The rusher who applies the function to the raw expression is the one who gets these wrong.
Mistake 3: Assuming The Derivative Is 0 Everywhere
Where it slips in: differentiating the signum function.
Don't do this: write $\frac{d}{dx}\operatorname{sgn}(x) = 0$ for all $x$, including the origin.
The correct way: the derivative is $0$ only for $x \neq 0$; at $x = 0$ it is undefined, because the function is discontinuous there. Continuity is a prerequisite for differentiability, and the jump breaks it.
Conclusion
The signum function $\operatorname{sgn}(x)$ returns $-1$, $0$, or $+1$: the sign of $x$, with size discarded.
Its domain is all real numbers; its range is the three-element set ${-1, 0, 1}$.
The graph is two flat rays with a single dot at the origin and a jump of $2$ at $x = 0$.
It is odd, multiplicative, bounded, and discontinuous only at $x = 0$; its derivative is $0$ off the origin and undefined there.
The usual slips are mishandling $\operatorname{sgn}(0)$, applying the function before simplifying, and assuming the derivative is $0$ everywhere.
To go deeper into functions and their graphs with a teacher, explore Bhanzu's algebra tutor sessions, a high school math tutor for board-level work, or structured math classes online. Want a live trainer to walk through the signum graph and other piecewise functions? Book a free demo class with a Bhanzu trainer.
Read More
Absolute value function — the sibling function that keeps size and discards sign.
Floor and ceiling function — more step-style functions built from flat pieces and jumps.
Relations and functions — the framework of domain, range, and mappings.
Identity function — the simplest function of all, for contrast with the signum.
Modulus of a complex number — where the idea of "size without direction" extends beyond the real line.
Was this article helpful?
Your feedback helps us write better content