You are halfway through a problem. A \sqrt{a + b} shows up. Your pen, without asking permission, writes:

\sqrt{a + b} = \sqrt{a} + \sqrt{b}.

It looks so clean. The square-root symbol sits in front of a sum; you "distribute" it over the two terms; you walk away feeling productive. And you are wrong. Not slightly wrong, not wrong in a corner case — wrong almost every single time this identity gets written. Examiners have been collecting this mistake for fifty years; it sits comfortably in the top five algebra errors on standardised tests, right next to (a+b)^2 = a^2 + b^2 (see the sibling article on algebra tiles).

This article is about why the move is illegal, what distribution rule you do have for square roots, and how to recognise the same mistake in all its disguises.

Test it with numbers

You do not need a theorem to catch this one. Pick a = 9 and b = 16 — two perfect squares so the arithmetic is clean.

\sqrt{9 + 16} = \sqrt{25} = 5.
\sqrt{9} + \sqrt{16} = 3 + 4 = 7.
5 \neq 7.

The identity fails on its very first numerical test. In any proof, in any exam paper, that single counterexample is enough to mark the statement wrong. You do not need to explain why — you have produced two numbers for which the two sides disagree, and "for all a and b" is dead.

Any time you are tempted to write \sqrt{a + b} = \sqrt{a} + \sqrt{b}, run this check in your head. Pick easy numbers, compute both sides, see them disagree, and put the pen down.

Why it fails — the square root is nonlinear

A function f is called linear when it satisfies

f(a + b) = f(a) + f(b)

for every a and every b. Linear functions are the only ones that "distribute over addition" — that is what linearity means. The classic examples are f(x) = 3x and f(x) = x/5: their graphs are straight lines through the origin, and you can pull apart a sum inside the function.

The square root f(x) = \sqrt{x} is not linear. Its graph is not a straight line — it is a curve that rises steeply near 0 and flattens out as x grows. Curving functions do not distribute over addition; the whole point of the curve is that the output at a + b is not the same as the outputs at a and b added together.

The square-root curve with three points showing that √25 is not √9 + √16A coordinate plane showing the graph of y equals square root of x from zero to thirty. Three points are marked: nine comma three, sixteen comma four, and twenty-five comma five. A dashed horizontal line at y equals seven shows where the sum three plus four would land, clearly above the five on the curve.xy091625123457(9, 3)(16, 4)(25, 5)y = 7 (the wrong "sum")gap of 2
The curve $y = \sqrt{x}$ passes through $(9, 3)$, $(16, 4)$, and $(25, 5)$. The value you get by adding the two ordinates — $3 + 4 = 7$ — sits well above the actual curve. The curve's value at $25$ is only $5$. The gap between $7$ and $5$ is the whole reason $\sqrt{a + b}$ is not $\sqrt{a} + \sqrt{b}$.

If the square-root graph were a straight line through the origin, the identity would hold. It is not, so it does not.

What DOES distribute over √

The square root is perfectly well-behaved over multiplication. For any a \geq 0 and b \geq 0,

\sqrt{a \cdot b} = \sqrt{a} \cdot \sqrt{b}.

You can check this with the same numbers. \sqrt{9 \cdot 16} = \sqrt{144} = 12, and \sqrt{9} \cdot \sqrt{16} = 3 \cdot 4 = 12. They agree. They will always agree, for every non-negative a and b — because squaring both sides gives a \cdot b = a \cdot b, which is obviously true.

So the rule is not "no distribution at all." The rule is: the square root distributes over multiplication, never over addition. Memorise the distinction. It is one of the sharpest lines in algebra.

Why multiplication works and addition doesn't

Watch what happens when you square each "candidate identity" and see which one is true.

The multiplication case. Is \sqrt{a} \cdot \sqrt{b} equal to \sqrt{a b}? Square both sides and see.

\left(\sqrt{a} \cdot \sqrt{b}\right)^2 = \left(\sqrt{a}\right)^2 \cdot \left(\sqrt{b}\right)^2 = a \cdot b.

The left side squared is a b. The right side squared is also a b. They match, so the identity is true.

The addition case. Is \sqrt{a} + \sqrt{b} equal to \sqrt{a + b}? Square the left side.

\left(\sqrt{a} + \sqrt{b}\right)^2 = \left(\sqrt{a}\right)^2 + 2 \sqrt{a}\sqrt{b} + \left(\sqrt{b}\right)^2 = a + 2\sqrt{ab} + b.

This is a + b + 2\sqrt{ab}, not a + b. The two sides disagree by exactly 2\sqrt{ab} — a leftover term that refuses to go away.

That 2\sqrt{ab} is the missing piece. It is what makes \sqrt{a + b} genuinely different from \sqrt{a} + \sqrt{b}, and it is structurally identical to the 2ab that goes missing in (a+b)^2 = a^2 + b^2. The cross-term is always where the mistake lives.

Where the identity would be true

The only way to kill the 2\sqrt{ab} term is to make it zero. That means \sqrt{ab} = 0, which means a b = 0, which means a = 0 or b = 0.

So the identity \sqrt{a + b} = \sqrt{a} + \sqrt{b} is true only in the trivial case where one of the numbers is zero: \sqrt{0 + 5} = \sqrt{0} + \sqrt{5} is fine because \sqrt{0} = 0. For any two positive a and b, the identity fails. There is no positive pair where it holds. None.

Same family of distribution errors

The "distribute over addition" bug is not specific to square roots. It infects every nonlinear operation a student meets. Here is the full rogues' gallery.

The pattern is unmissable. Nonlinear operations never distribute over addition. Only linear operations — scalar multiplication, the identity function, differentiation, integration — distribute over addition. Every time you see a function wrapped around a sum, stop and ask: is this function linear? If not, you cannot pull it apart.

What to do when you actually want to split √(a+b)

In general, you cannot. The expression \sqrt{a + b} is often stuck the way it is, and the best move is to leave it alone and work around it. But three techniques sometimes rescue you.

Factor first. If the expression inside the root has a common factor, pull it out and then take its square root.

\sqrt{4 + 4x} = \sqrt{4(1 + x)} = 2\sqrt{1 + x}.

You have not split the 1 + x; you have only extracted the 4. The \sqrt{1 + x} part stays as one piece.

Complete a square. If the sum inside is secretly a perfect square, rewrite it and take the root exactly.

\sqrt{x^2 + 2x + 1} = \sqrt{(x + 1)^2} = |x + 1|.

Here the addition inside was never a generic sum — it was the expansion of (x+1)^2, which is why the root simplifies. You did not distribute; you recognised a square.

Taylor approximate (for small terms). In physics and advanced calculations, when one piece is tiny, you use the expansion

\sqrt{1 + \varepsilon} \approx 1 + \frac{\varepsilon}{2} \quad \text{for small } \varepsilon.

This is an approximation, not an identity, and it only works when \varepsilon is close to zero. You will see it often in JEE physics (relativistic corrections, pendulum amplitude corrections) but it is not a licence to split general square roots.

When the identity "looks" to work

Watch out for two near-misses that can fool you into thinking the general rule holds.

Zero sneaks in. \sqrt{1 + 0} = \sqrt{1} = 1, and \sqrt{1} + \sqrt{0} = 1 + 0 = 1. The identity agrees — but only because one operand is zero, which is the trivial case established above. Do not generalise from this.

A hidden perfect square. \sqrt{a^2 + 2ab + b^2} = \sqrt{(a + b)^2} = |a + b|, which equals a + b when both are non-negative. This looks like \sqrt{a^2} + \sqrt{b^2} plus something distributed — but it is not. The reason the root came out cleanly is that the sum inside was already a perfect square. You did not distribute anything; you recognised a square. That is a completely different move.

Recognition drill

Which of the following can you simplify, and which are trapped as written? Go through them in your head before reading the answers.

If any of these tricked you, run the numerical test again. Pick small values, evaluate both sides, and watch where the identity breaks.

The one-line takeaway

The square root does not distribute over addition. It distributes over multiplication. That distinction is a single misconception that costs students marks in every algebra exam and survives decades of classrooms. The defence is cheap: every time you are tempted to write \sqrt{a + b} = \sqrt{a} + \sqrt{b}, plug in a = 9 and b = 16, watch 5 disagree with 7, and put the identity back in the bin where it belongs.