Extraneous and Lost Roots

Start with something simple. The equation \sqrt{x} = -3 has no solution — a square root is never negative. But watch what happens if you square both sides: you get x = 9. Plug x = 9 back in: \sqrt{9} = 3 \neq -3. The number 9 looks like a solution if you only stare at the squared version, but it fails the original equation completely. Where did it come from?

Or try this in the other direction. The equation x(x - 2) = 0 has two solutions: x = 0 and x = 2. Now divide both sides by x: you get x - 2 = 0, so x = 2. The root x = 0 has vanished. It was a perfectly good solution, and a single algebraic step made it disappear.

These two phenomena — solutions appearing out of nowhere, and solutions silently vanishing — are not rare accidents. They happen in predictable places, for clear reasons. Once you see the mechanism, you stop being surprised and start being careful.

How extraneous roots arise

An extraneous root is a number that satisfies the transformed equation but not the original one. It is created when an algebraic step maps two different equations into the same equation — the step is not reversible, so information is lost in the forward direction, and a fake solution slips through in the backward direction.

Three moves are the usual culprits.

Squaring both sides

If a = b, then a^2 = b^2. That is true. But the reverse is not: a^2 = b^2 only means a = b or a = -b. Squaring destroys the sign, so when you square an equation, you are also accepting solutions where one side equals the negative of the other.

Take \sqrt{x + 3} = x - 3. Squaring gives x + 3 = (x - 3)^2 = x^2 - 6x + 9, which rearranges to x^2 - 7x + 6 = 0, factoring as (x - 1)(x - 6) = 0. Two candidates: x = 1 and x = 6.

Check x = 1: left side = \sqrt{4} = 2, right side = 1 - 3 = -2. Not equal. Extraneous.

Check x = 6: left side = \sqrt{9} = 3, right side = 6 - 3 = 3. Equal. Genuine.

The squaring step accepted x = 1 because at that value \sqrt{x+3} equals the negative of x - 3. The squared equation cannot tell the difference.

Multiplying both sides by an expression that can be zero

If you multiply both sides of an equation by (x - 5), you create a new equation that is satisfied whenever x = 5 — regardless of whether the original was. You have grafted an extra root onto the equation.

Consider \dfrac{x}{x - 5} = \dfrac{5}{x - 5}. The natural first move is to multiply both sides by (x - 5), giving x = 5. But plug x = 5 back in: both denominators become zero, so the original equation is undefined at x = 5. There is no solution at all — the candidate x = 5 is extraneous.

Cross-multiplying a rational equation

This is a special case of the previous move. Any time you clear denominators in a rational equation, you are implicitly multiplying by factors that might be zero. The candidates you get at the end must be tested against the original denominators.

How roots are lost

Lost roots are the mirror image: a step narrows the equation, throwing away solutions that were genuinely there.

Dividing both sides by an expression containing the variable

If x \cdot f(x) = x \cdot g(x) and you divide both sides by x, you get f(x) = g(x) — but you have silently assumed x \neq 0. If x = 0 was a solution of the original, it is now gone.

Take x^2 = 4x. Dividing both sides by x gives x = 4. But the original factors as x(x - 4) = 0, giving x = 0 and x = 4. The root x = 0 was lost.

The safe alternative: move everything to one side and factor. x^2 - 4x = 0, so x(x - 4) = 0, and both roots survive.

Taking a single branch of a square root

When you solve x^2 = 9 by writing x = 3, you have lost x = -3. The correct move is x = \pm 3. This is so common that it barely feels like "losing a root," but the mechanism is the same — you chose one of two valid branches and forgot the other.

Imposing domain restrictions prematurely

If an equation involves \sqrt{x - 1}, you know x \ge 1. But some students write that restriction and then carry it through in a way that discards a valid root of a different piece of the equation. Keep the restriction for checking, not for solving — find all algebraic candidates first, then filter at the end.

Checking solutions

The single most reliable habit in algebra is this: always substitute every candidate back into the original equation. Not the simplified version. Not the version after you squared. The original, exactly as it was written before you touched it.

The check has two jobs.

1. Catch extraneous roots. A candidate that fails the original equation is extraneous — discard it.
2. Confirm genuine roots. A candidate that passes is a real solution.

For equations involving square roots, a correct check also reveals why the extraneous root appeared — you will find that the two sides have equal magnitude but opposite sign, because squaring erased the sign difference.

For rational equations, the check is even simpler: plug the candidate into each denominator. If any denominator hits zero, the candidate is extraneous, regardless of what the numerator does.

Here is an interactive test. The equation is \sqrt{x + 3} = x - 3. Drag the red point along the x-axis to any candidate value. The two readouts show the left side \sqrt{x+3} and the right side x - 3. A genuine root is a position where both readouts agree.

Avoiding errors

Knowing why extraneous roots appear and why roots get lost is the best defence, but a few mechanical habits make the process almost automatic.

Before you solve:

• Note every domain restriction. If the equation has \sqrt{f(x)}, write f(x) \ge 0. If it has \dfrac{1}{g(x)}, write g(x) \neq 0. Keep these restrictions visible while you work.

While you solve:

• Never divide both sides by an expression containing the variable. Move everything to one side and factor instead. This prevents lost roots.
• When you must square (or raise to any even power), accept that the step is non-reversible. Mark the candidates as "needs checking."
• If you multiply both sides by a variable expression to clear denominators, note which values would make that expression zero. Those values are automatically suspect.

After you solve:

• Substitute every candidate into the original equation. Every single one, no shortcuts.
• For radical equations, verify that both sides have the same sign as well as the same magnitude.
• For rational equations, verify that no denominator is zero.

The pattern is: find all candidates, then filter. The algebraic machinery finds candidates; the checking step filters out the fakes and confirms the real ones.

Worked examples

Common confusions

• "If my algebra gives two answers, both must be correct." Not if you squared or multiplied by a variable expression. The algebra finds candidates; checking confirms which are real.

• "I only need to check when there's a square root." Any non-reversible step can create extraneous roots. Clearing denominators in rational equations is just as dangerous as squaring radical equations. Absolute value equations, too — removing the absolute value bars involves a case split, and skipping a case can both create extraneous roots and lose real ones.

• "Dividing both sides by x is fine because x is just a number." It is a number you don't know yet. If it turns out to be zero, dividing by it was dividing by zero — undefined and destructive. The safe move is always to factor, not to divide.

• "An equation with no solution must be wrong." No — the equation \dfrac{x^2-9}{x-3} = 6 is perfectly well-posed. It just happens that its solution set is empty. The algebra produces a candidate that the domain excludes. That is a valid mathematical outcome, not an error.

• "Extraneous roots come from mistakes." They come from correct algebraic steps that happen to be non-reversible. Squaring both sides of \sqrt{x} = -3 is a legal algebraic operation. The result x = 9 is a correct solution of the squared equation. It just is not a solution of the original equation. The step was correct; the conclusion would be wrong only if you forgot to check.

• "If the check fails, I must have done the algebra wrong." Not necessarily. If the check fails, the candidate is extraneous — it was never a real solution. A failed check is the system working, not breaking. Go back and redo the algebra only if every candidate fails and you expected at least one to survive.

Going deeper

The underlying idea: non-invertible functions

Every extraneous-root problem has the same abstract shape. You start with an equation f(x) = g(x). You apply some function h to both sides to get h(f(x)) = h(g(x)). If h is injective (one-to-one) — meaning different inputs always give different outputs — then the new equation has exactly the same solutions as the old one. No extraneous roots can appear.

But if h is not injective, different inputs can map to the same output. Squaring is the classic example: h(t) = t^2 sends both 3 and -3 to 9. So the equation h(f(x)) = h(g(x)) is satisfied whenever f(x) = g(x) or f(x) = -g(x). The second case is the source of the extraneous roots.

Lost roots work the same way in reverse. Dividing both sides by x is the same as applying h(t) = t/x — but this function is undefined at x = 0, so any solution at x = 0 is thrown away. The function h was not defined on the full domain of the original equation, and the missing piece of the domain took a root with it.

The general principle: an algebraic step that applies an injective function on the full domain of the equation preserves the solution set exactly. Any deviation — h is not injective, or h is not defined on the whole domain — can either add extraneous roots or lose genuine ones.

How competitive exams test this

JEE and other competitive papers rarely ask "find the extraneous root" directly. Instead, they embed the trap inside a larger problem. A common pattern: an equation has two candidate solutions, and the answer options list both individual values plus their sum. Students who skip the check choose the sum; students who check find that one candidate is extraneous and choose the single survivor. The entire question is testing whether you habitually check.

Another pattern: a problem asks "how many solutions does this equation have?" The algebra gives three candidates, but two are extraneous. The answer is one, and students who trust the algebra without checking answer three. These are free marks for anyone who has internalised the checking habit.

Connection to equivalence transformations

In formal algebra, a step that preserves the solution set is called an equivalence transformation. Adding the same expression to both sides (when defined) is an equivalence transformation. Multiplying both sides by a non-zero constant is an equivalence transformation. Squaring is not — it is an implication transformation, meaning the new equation is implied by the old one but does not imply it back. The arrow goes one way, and the backward direction is where the extraneous roots hide.

When you write \Leftrightarrow (double arrow) between two lines of algebra, you are claiming equivalence — the two equations have exactly the same solutions. When you write \Rightarrow (single arrow), you are claiming only implication — the new equation may have more solutions. The discipline of choosing the right arrow is what experienced problem-solvers use to keep track of where extraneous roots might appear.

Where this leads next

The checking habit you built here applies in every equation-solving context that follows.

• Linear Equations in One Variable — the simplest equations, where every step is reversible and extraneous roots cannot appear. A useful contrast.
• Quadratic Equations — Introduction — the discriminant tells you how many roots to expect, which makes the check faster: if the discriminant says two and you found two, both should survive.
• Absolute Value Equations — case splits create the same "widening" effect as squaring. Each case must be checked against the condition that defined it.
• Radicals and Rational Exponents — deeper work with square and cube roots, where squaring is the main tool and extraneous roots are the main hazard.
• Polynomial Equations — factoring over higher-degree polynomials, where the "factor instead of dividing" principle becomes even more important.