You have memorised a(b + c) = ab + ac and watched it work every time. So when you meet (b + c)^2, your brain reaches for the same pattern: surely (b + c)^2 = b^2 + c^2? You are not the first student to write this on a test, and you will not be the last. But the two sides aren't equal — there is always an extra middle term — and the reason isn't a technicality. It comes from what exponentiation actually is.

The wrong move, tested on numbers

Try b = 3, c = 4 and compute both sides of the imagined identity (b + c)^2 = b^2 + c^2.

Off by 24. That 24 is 2 \times 3 \times 4 = 2bc — a clue we will return to in a moment. So the distributive-looking move breaks, and it breaks by a specific, predictable amount.

Why multiplication distributes: it is a single action

The rule a(b + c) = ab + ac is a clean distribution because multiplication by a means scaling — you stretch the quantity (b + c) by a factor of a. Scaling acts on each piece independently. If you scale a rectangle of width b + c by factor a, you scale the b-part and the c-part separately and add. One copy of a touches each piece.

Why this distributes cleanly: there is exactly one multiplication happening. a is applied once to b and once to c. Nothing interacts with anything else. The rule says "apply the same a to each piece independently," and that's exactly what scaling does.

Why exponentiation doesn't: squaring is two multiplications

Now expand (b + c)^2 without being lazy. The exponent 2 means "multiply the thing by itself."

(b + c)^2 = (b + c)(b + c).

Now you have a product of two sums, not a single distribution. Use the real distributive law on each factor:

(b + c)(b + c) = b(b + c) + c(b + c)
= b^2 + bc + cb + c^2
= b^2 + 2bc + c^2.

The extra 2bc is the cross-term. It shows up because when you multiply the first (b + c) by the second (b + c), every term of the first meets every term of the second. There are four pairings: b \cdot b, b \cdot c, c \cdot b, c \cdot c. The two middle pairings collapse to 2bc, which is the piece the lazy formula misses.

So the correct identity — the one you actually need to memorise — is

(b + c)^2 = b^2 + 2bc + c^2.

The b^2 and c^2 are there. It is the 2bc that explains why distribution doesn't work the way you wanted.

The picture — why the cross-term has to exist

This becomes vivid geometrically. (b + c)^2 is the area of a square of side b + c. Cut that square into four regions along a vertical and horizontal line at distance b from one corner.

Geometric expansion of b plus c squared into four regions A large square of side b plus c, divided by one horizontal and one vertical line into four smaller regions. The top-left region is a square of side b with area b squared. The bottom-right region is a square of side c with area c squared. The top-right region is a rectangle of width c and height b with area b times c. The bottom-left region is a rectangle of width b and height c with area b times c. The total area equals b squared plus two b c plus c squared. b² b·c b·c c² b c b c (b + c)² = b² + 2bc + c²
The square of side $b + c$ decomposes into four pieces. Two of them — the $b \times b$ and $c \times c$ squares — account for $b^2 + c^2$. The other two are rectangles of area $b \times c$ each, accounting for the $2bc$ cross-term. Those two rectangles are exactly what the naive rule $(b+c)^2 = b^2 + c^2$ forgets. You can see them in the picture; there is nowhere for them to hide.

The four pieces total b^2 + bc + bc + c^2 = b^2 + 2bc + c^2. Two of those pieces are the cross-term rectangles, and they are half the decomposition by area in many cases. If you write b^2 + c^2, you are literally deleting two rectangles from the picture. The identity has to include them or it isn't describing the same shape.

A deeper reason: exponents count, they don't scale

The cleanest way to state the distinction is this.

So distributivity is about scaling (one-at-a-time). Squaring is about pairing (each-with-each). The two aren't the same operation and don't obey the same rule.

What about (bc)^2? Does that distribute?

Funnily enough, exponentiation does distribute — but over multiplication, not addition:

(bc)^2 = b^2 c^2.

Check: (3 \cdot 4)^2 = 12^2 = 144, and 3^2 \cdot 4^2 = 9 \cdot 16 = 144. Match. This works because squaring bc gives bcbc = b \cdot b \cdot c \cdot c = b^2 c^2 — the b's collect together, the c's collect together, and no cross-terms appear because multiplication commutes freely.

So the clean rule is:

Exponents distribute over multiplication but not over addition.

That is the single sentence to remember. (bc)^n = b^n c^n for any n, but (b + c)^n almost never equals b^n + c^n.

The full pattern: binomial theorem

For higher powers, the cross-terms multiply. The full expansion of (b + c)^n is the binomial theorem:

(b + c)^n = \sum_{k=0}^{n} \binom{n}{k} b^{n-k} c^k.

For n = 2 the coefficients are 1, 2, 1 — giving b^2 + 2bc + c^2. For n = 3 they are 1, 3, 3, 1 — giving b^3 + 3b^2c + 3bc^2 + c^3. The endpoints (b^n and c^n) are the only survivors of the naive rule; everything in the middle is cross-term.

This is why (b + c)^n = b^n + c^n is wrong in a predictable way: it drops every middle term of the binomial expansion.

A fast sanity check to kill the bad move forever

If you ever feel the urge to write (b + c)^2 = b^2 + c^2 on an exam, do this ten-second test.

Pick b = 1 and c = 1.

Off by a factor of 2. Wrong. Sanity check done; back up and expand properly. This one-line test catches the mistake every time, even when you are half-asleep.

Summary

Related: Operations and Properties · Distributive Property as an Area: a(b + c) Is Literally Two Rectangles · Algebraic Identities · Exponents and Powers · Binomial Theorem for Positive Integer