How To Show If A Function Is One To One

You’re staring at a problem set. The prompt asks you to prove a function is one-to-one. Your stomach does that little drop thing. You know the definition — sort of — but turning that definition into a clean, logical argument? That’s where the wheels fall off.

Been there. Still, the good news? It’s the standard rite of passage in discrete math, real analysis, and basically any proof-based course after calculus. There are really only a handful of moves. Once you recognize the pattern, it stops feeling like magic and starts looking like algebra with a logic backbone.

Let’s walk through it. No fluff. Just the moves that actually work.

What Is a One-to-One Function

A function $f: A \to B$ is one-to-one (or injective*) if distinct inputs always produce distinct outputs. Formally: for every $x_1, x_2 \in A$, if $f(x_1) = f(x_2)$, then $x_1 = x_2$.

That’s the definition. Here's the thing — memorize it. Say it in your sleep.

The contrapositive is often easier to work with: if $x_1 \neq x_2$, then $f(x_1) \neq f(x_2)$. Consider this: same logical content, different flavor. Pick whichever feels more natural for the problem at hand.

Visually, it means no horizontal line hits the graph more than once. That’s the horizontal line test. It’s a great intuition check, but it’s not a proof — unless you’re in a context where graphical reasoning is explicitly allowed (rare in upper-division math).

The Core Idea: Collisions Are Forbidden

Think of it like a hotel. The domain is the set of guests. In real terms, the codomain is the set of rooms. A function assigns each guest to a room. One-to-one means no two guests share a room. Simple as that.

If you find even one pair of distinct inputs mapping to the same output, the function fails. Just one counterexample kills it.

Why It Matters

Injectivity isn’t just a vocabulary word. It’s the gateway to invertibility. A function has a left inverse if and only if it’s injective. It has a two-sided inverse (a true inverse function) if and only if it’s bijective* — both one-to-one and onto.

In linear algebra, a linear transformation is injective exactly when its kernel is trivial. Consider this: in group theory, a homomorphism is injective exactly when its kernel is the identity. In topology, embeddings are injective continuous maps with a continuous inverse on their image.

The concept keeps showing up under different names. Kernel trivial. That said, embedding. On top of that, section. Monomorphism. Same structural idea: no collapsing.

If you can’t prove a function is one-to-one, you can’t prove it’s invertible. You can’t safely “cancel” it from the left in equations. You lose a massive toolkit.

How to Prove It: The Standard Methods

There isn’t one universal technique. The right approach depends on how the function is given — formula, graph, table, set of ordered pairs, or abstract definition. Here are the main playbooks.

Direct Proof from the Definition

This is the default. Assume $f(x_1) = f(x_2)$. Manipulate until you get $x_1 = x_2$.

Example: $f: \mathbb{R} \to \mathbb{R}$ defined by $f(x) = 3x - 7$.

Assume $f(x_1) = f(x_2)$.
Here's the thing — then $3x_1 - 7 = 3x_2 - 7$. Consider this: add 7: $3x_1 = 3x_2$. Because of that, divide by 3: $x_1 = x_2$. Done.

Clean. Day to day, linear functions with nonzero slope are always injective. The algebra forces the inputs to match.

Example: $f: \mathbb{R} \to \mathbb{R}$ defined by $f(x) = x^3$.

Assume $x_1^3 = x_2^3$.
Take cube roots (valid on all reals): $x_1 = x_2$.
Done.

Notice: $f(x) = x^2$ fails on $\mathbb{R}$ because $(-2)^2 = 2^2$ but $-2 \neq 2$. But restrict the domain to $[0, \infty)$ and it becomes injective. Domain matters. Always check the domain.

Using Derivatives (Calculus Approach)

If $f$ is differentiable on an interval and $f'(x) > 0$ everywhere (or $f'(x) < 0$ everywhere), then $f$ is strictly monotone, hence injective.

Example: $f(x) = x^3 + x$ on $\mathbb{R}$.
$f'(x) = 3x^2 + 1 > 0$ for all $x$. Strictly increasing. One-to-one.

This is powerful but has caveats. The derivative must not change sign. Because of that, $f(x) = x^3$ has $f'(0) = 0$ but is still injective — the derivative test is sufficient, not necessary. And it only works on intervals (connected domains). If the domain is disconnected, monotonicity on each piece doesn’t guarantee global injectivity.

Algebraic Manipulation for Rational / Radical Functions

Example: $f(x) = \frac{2x+1}{x-3}$, domain $\mathbb{R} \setminus {3}$.

Assume $\frac{2x_1+1}{x_1-3} = \frac{2x_2+1}{x_2-3}$.
Rearrange: $7x_2 = 7x_1$.
Cross-multiply: $(2x_1+1)(x_2-3) = (2x_2+1)(x_1-3)$.
Plus, expand: $2x_1x_2 - 6x_1 + x_2 - 3 = 2x_1x_2 - 6x_2 + x_1 - 3$. Now, cancel $2x_1x_2$ and $-3$: $-6x_1 + x_2 = -6x_2 + x_1$. So $x_1 = x_2$.

The algebra gets messy but it’s just symbol pushing. Stay organized. Write each step on its own line.

Piecewise Functions

Break it into cases. The function might be defined differently on different subsets of the domain. Even so, you have to check:

Injectivity within* each piece. Worth adding: 2. No collisions across* pieces.

Example:
$f(x) = \begin{cases} x+1 & x < 0 \ x^2 & x \ge 0 \end{cases}$

On $(-\infty, 0)$: $x+1$ is linear, slope 1, injective.
On $[0, \infty)$: $x^2$ is strictly increasing (derivative $2x \ge 0$, zero only at 0), injective.
Cross-piece: outputs from first piece are ${content}lt; 1$. Outputs from second piece are $\ge 0$. Overlap on $[0,1)$. Uh oh.
$f(-0.5) = 0.5$ and $f(\sqrt{0.5}) = 0.5$. Distinct inputs, same output. Not injective.

Always check the boundaries.

Trigonometric, Exponential, and Logarithmic Families

A lot of the “textbook” functions you meet in calculus are automatically injective on suitable intervals, but you still have to check the domain.

Sine and Cosine

Both $\sin x$ and $\cos x$ are periodic, so on $\mathbb{R}$ they’re not injective.
On the flip side, if you restrict the domain to a half‑period, they become one‑to‑one:

$\sin x$ on $[-\tfrac{\pi}{2},\tfrac{\pi}{2}]$ is strictly increasing.
$\cos x$ on $[0,\pi]$ is strictly decreasing.

Proof sketch for $\sin$ on $[-\tfrac{\pi}{2},\tfrac{\pi}{2}]$:
Assume $\sin x_1 = \sin x_2$ with $x_1,x_2\in[-\tfrac{\pi}{2},\tfrac{\pi}{2}]$.
Use the identity $\sin x_1 - \sin x_2 = 2\cos\bigl(\tfrac{x_1+x_2}{2}\bigr)\sin\bigl(\tfrac{x_1-x_2}{2}\bigr)=0$.
Since $\cos\bigl(\tfrac{x_1+x_2}{2}\bigr)\neq 0$ on that interval, we must have $\sin\bigl(\tfrac{x_1-x_2}{2}\bigr)=0$, forcing $x_1-x_2=0$, i.e. $x_1=x_2$.

Exponential

The exponential function $e^x$ is strictly increasing on all of $\mathbb{R}$, so it’s injective without restriction.
If you take $f(x)=a^x$ with $a>0$, $a\neq1$, the same monotonicity argument applies: $f'(x)=a^x\ln a$ never changes sign.

Logarithm

$\ln x$ is defined only on $(0,\infty)$ and is strictly increasing there.
Thus, if you ever see a function of the form $\ln(g(x))$, the injectivity of $f$ hinges on two things:

$g$ must map its domain into $(0,\infty)$.
$g$ itself must be injective on the domain you care about.

Take this: $f(x)=\ln(x^2+1)$ is injective on $\mathbb{R}$ because $x^2+1$ is always positive and strictly increasing on $[0,\infty)$, but it’s not injective on all of $\mathbb{R}$ because $x^2+1$ is even.

General Strategies That Work Everywhere

Algebraic Re‑arrangement – When you can solve the equation $f(x_1)=f(x_2)$ explicitly for $x_1$ in terms of $x_2$, you’re done.
Typical for rational, polynomial, and many algebraic functions.*

If you found this helpful, you might also enjoy which of the following drugs is not a hallucinogen or how many volts is 1 joule.

If you found this helpful, you might also enjoy which of the following drugs is not a hallucinogen or how many volts is 1 joule.
Monotonicity via Derivatives – If you can show $f'(x)$ never changes sign on a connected set, injectivity follows.
Be careful with critical points where $f'=0$—the function can still be injective, but the derivative test is only a sufficient condition.*
Inverse Construction – If you can write an explicit inverse $f^{-1}$ that is well‑defined on the codomain, injectivity is guaranteed.
Often used for trigonometric and exponential functions on restricted domains.*
Piecewise Analysis – Break the domain into pieces where the function behaves nicely.
Always check cross‑piece collisions; the most subtle failures happen at the boundaries.*
Domain Restriction – A function that is not injective on its natural domain memorabilia can become injective after trimming the domain.
Example: $x^2$ on $\mathbb{R}$ is not injective, but on $[0,\infty)$ it is.*

Common Pitfalls to Watch For

Assuming Continuity Implies Injectivity – A continuous function can still “fold” over itself (e.g., $\cos x$).
Ignoring Domain/Range Constraints – $\ln(x)$ is only defined for $x>0$; trying to prove injectivity on $\mathbb{R}$ is meaningless.
Overlooking Boundary Points – In piecewise functions, the values at the junctions can create duplicates that you might miss if you only look at the interior of each piece.
Misreading Derivative Sign – A derivative that is zero at isolated points does

Misreading Derivative Sign

A derivative that vanishes at isolated points does not automatically destroy injectivity.
And consider (f(x)=x^{3}). Its derivative (f'(x)=3x^{2}) is zero at (x=0), yet (f) is strictly increasing on (\mathbb{R}) and therefore injective. The key is to examine the sign of the derivative on each interval of the domain: if (f') never changes sign (even if it hits zero at a point), the function remains monotone and hence one‑to‑one.

Most people don't realize how important this is.

A common mistake is to look only at the existence of critical points and conclude that the function “folds.On the flip side, ” The correct approach is to check whether the derivative can switch from positive to negative (or vice‑versa) across a critical point. If it does not, the function can still be injective despite isolated zeros of (f').

Overlooking Asymptotic Behavior

Even when a function is monotone on its entire domain, one must verify that its range does not repeat values at infinity. Here's one way to look at it: the function

[ g(x)=\arctan x \quad (x\in\mathbb{R}) ]

is strictly increasing and injective, but its codomain (\bigl(-\tfrac{\pi}{2},\tfrac{\pi}{2}\bigr)) is bounded. If one mistakenly treats (\arctan x) as a map onto (\mathbb{R}), the claim of injectivity would be vacuous. Always match the range to the intended codomain; injectivity is a property of the mapping, not of the underlying set.

Confusing Local Injectivity with Global Injectivity

A function may be locally injective (i.e., injective on a small neighbourhood of each point) without being globally injective.

[ h(x)=\sin x \quad\text{on }[0,2\pi]. ]

On any interval of length less than (\pi) the sine function is monotone, but on the whole interval ([0,2\pi]) it attains the same value at (x=\tfrac{\pi}{2}) and (x=\tfrac{3\pi}{2}). That said, when you are asked to prove injectivity, you must decide whether the domain is the maximal one or a restricted one. If the domain is not specified, assume the natural domain and be prepared to restrict it if needed.

Ignoring the Role of Symmetry

Even‑odd symmetry can hide duplicate values. To give you an idea, the function

[ p(x)=\frac{x^{2}}{1+|x|} ]

is not injective on (\mathbb{R}) because (p(-x)=p(x)). The presence of an absolute value or an even power often signals a potential symmetry that must be examined. A quick check—compare (p(x)) with (p(-x))—can save a lot of effort later.

A Unified Checklist for Proving Injectivity

Domain and Codomain – Verify that the domain is well‑defined and that the intended codomain contains all possible outputs.
Algebraic Simplification – Try to solve (f(x_{1})=f(x_{2})) for (x_{1}) in terms of (x_{2}). If the only solution is (x_{1}=x_{2}), you are done.
Monotonicity Test – Compute (f') (or a suitable difference quotient) and confirm that it never changes sign on a connected component of the domain.
Inverse Construction – If you can write an explicit inverse (f^{-1}) on the codomain, injectivity follows automatically.
Piecewise Examination – Split the domain where the formula changes, check monotonicity on each piece

and verify that the images of distinct pieces do not overlap.
7. 6. Day to day, Boundary and Asymptotic Analysis – Confirm that limits at endpoints or infinities do not produce duplicate values already attained inside the domain. Symmetry Check – Test whether (f(x)=f(-x)) or (f(x)=f(c-x)) for some constant (c); if so, restrict the domain or prove the symmetry forces (x_1=x_2).

Worked Example: A Rational Function with Absolute Value

Consider

[ f(x)=\frac{x|x-2|}{x^2+1},\qquad x\in\mathbb{R}. ]

Step 1 – Domain and Codomain. The denominator never vanishes, so the domain is (\mathbb{R}). The codomain is (\mathbb{R}) (we will see the range is actually bounded).

Step 2 – Piecewise Definition.
[ f(x)=\begin{cases} \dfrac{x(2-x)}{x^2+1}, & x<2,\[6pt] \dfrac{x(x-2)}{x^2+1}, & x\ge 2. \end{cases} ]

Step 3 – Monotonicity on Each Piece.
For (x<2): (f'(x)=\frac{-x^2+4x+2}{(x^2+1)^2}). The numerator has roots (2\pm\sqrt{6}); only (2-\sqrt{6}\approx -0.45) lies in ((-\infty,2)). The derivative changes sign there, so (f) is not monotone on the whole piece.

For (x\ge 2): (f'(x)=\frac{x^2-4x+2}{(x^2+1)^2}). The numerator is positive for (x>2+\sqrt{2}) and negative on ([2,2+\sqrt{2})). Again, not monotone.

Step 4 – Algebraic Approach. Solve (f(x_1)=f(x_2)). Cross-multiplying and factoring yields
[ (x_1-x_2)\bigl[(x_1x_2-1)(|x_1-2|+|x_2-2|)+ (x_1+x_2)(|x_1-2|-|x_2-2|)\bigr]=0. ]
The second factor vanishes only when (x_1=x_2) (a tedious but straightforward case analysis on the signs of (x_i-2) confirms this). Hence (f) is injective despite lacking global monotonicity.

When Calculus Fails: Discrete and Non‑Differentiable Settings

The derivative test is unavailable for functions on discrete domains, functions with corners, or maps between abstract sets. g.Think about it: in those cases fall back on the definition:
[ \forall x_1,x_2\in D,\quad f(x_1)=f(x_2)\implies x_1=x_2. In practice, ]
For integer sequences, use induction or number-theoretic properties (e. , strict growth of (n\mapsto 2^n+n^2)). For maps between finite sets, a pigeonhole-principle argument often suffices: if (|D|=|C|) and (f) is surjective, it is automatically injective.

Conclusion

Proving injectivity is rarely a one‑tool job. In practice, by systematically verifying the domain, dissecting the function into monotone segments, checking for hidden symmetries, and—when necessary—returning to the algebraic definition (f(x_1)=f(x_2)\Rightarrow x_1=x_2), you avoid the common pitfalls that turn a plausible claim into a flawed proof. A derivative sign chart gives a quick answer for smooth, connected domains, but piecewise definitions, symmetries, asymptotic plateaus, and non‑differentiable points all demand separate attention. Injectivity is a global property; treat it with the global scrutiny it deserves.