Visual walkthrough — Laws of exponents — review with real exponents
Before any algebra, let us agree on what a power even is — in a picture.
Step 1 — What a power looks like: a row of identical blocks
WHAT. The symbol (read " to the ") is shorthand for a row of identical copies of , all multiplied together. Here is the base (the thing being copied) and is the exponent (how many copies).
WHY a picture at all? Because the whole law we are chasing is nothing but counting blocks. If you can see the blocks, you never memorise — you just recount.
PICTURE. Below, each pale-yellow block is one copy of . The row labelled has three blocks; has two blocks.

We insist throughout. (Why not or ? Those get their own step — Step 7 — because they fail for different reasons.) Keep the block picture in your head — every step below is just sliding rows of blocks together.
Step 2 — Multiplying two powers = pushing two rows together
WHAT. Take one row of blocks () and a second row of blocks (). Multiplying them, , means placing the second row right after the first.
WHY this move? Multiplication just says "and then multiply by these too." So literally appends the -block row onto the -block row. No new rule invented — we are only rearranging blocks we already had.
PICTURE. The blue row () and the pink row () slide together into one long row.

- ::: the blue pile — copies
- ::: the pink pile — copies
- the space between them ::: the multiplication dot; it does nothing to the blocks, just joins the piles
Step 3 — Count the whole row: the exponents add
WHAT. How many blocks are in the joined row? Just count: blue ones, then pink ones, so blocks total. A single row of copies of is .
WHY the ""? Because we are counting a combined pile. Counting a group of things and then more things is addition — that is the only place the plus sign comes from. There is nothing deeper; the plus in the exponent is just "how many blocks altogether."
PICTURE. The same long row from Step 2, now with a single bracket underneath counting all the blocks as .

Step 4 — Force the law to survive at : why
WHAT. So far was at least — a row must have blocks. What if a row has zero blocks? We are not free to guess ; we demand that the product law from Step 3 keeps working, and see what it forces.
WHY force it? A rule with a hole in it (works for but breaks at ) is useless. We want one law, no exceptions. So we insist Step 3 stays true, plug in , and let the algebra tell us the only allowed value.
With that in hand, is the empty product (zero copies of ), so directly. The product law then simply confirms it:
- ::: an -row joined with an empty row
- ::: joining an empty row adds nothing, so the count is unchanged
- ::: divide both sides by ; since , that division is legal (never )
- ::: matches the empty-product value — the two viewpoints agree
PICTURE. An empty row (no blocks) tacked onto — the count stays , so the empty row must "weigh" .

Step 5 — Force negative exponents: why
WHAT. Now let the exponent go below zero. What could " blocks" mean? Again we do not guess — we force Step 3 to survive.
WHY. We want to append an -row to an -row and land on the empty row (). Whatever "un-does" blocks back to nothing is the meaning of .
- ::: a full row and its "anti-row"
- ::: they cancel to the empty row
- ::: from Step 4, the empty row is
- ::: the anti-row is whatever flips into its reciprocal so the product is
PICTURE. A row of blocks and a pink "anti-row" (drawn as a fraction bar) that cancels it down to a single unit block ().

Step 6 — Force fractional exponents: why
WHAT. Push the same "make the law survive" idea onto a fraction upstairs. We build this in two clicks: first the single root , then the general rational . We will need the companion power-of-a-power law , so let us earn that from blocks first.
WHY unpack power-of-a-power? Because is defined by demanding , and we may not use a law we have not seen in blocks.
PICTURE (still one long row). Below, three chunks of length are laid end to end (). The bracket counts all blocks as — no square, no area, just a longer row.

Now the two clicks:
Click 1 — single root. We want . So is the number whose -th power is — the -th root:
Click 2 — general rational . Write and apply power-of-a-power:
- ::: first take the -th root (one super-block)
- ::: then make a row of of them (repeat times)
- ::: root first, power second — a plain block-row of copies of
So a rational exponent like is fully block-defined: build the super-block , then lay of them in a row. This is the block-meaning of every decimal exponent used later. See Rational Exponents and Radicals for the full bridge.
Recall Why is a decimal like
a rational (fraction) exponent? A terminating decimal is just a fraction over a power of ten: , , . Each is with whole — so each already has the Step 6 block-meaning " copies of ." That is exactly why we may call them "already defined."
Step 7 — The degenerate bases: and (two different failures)
WHAT. Everything above quietly assumed . There are two excluded bases, and they misbehave for different reasons — so we separate them.
Case . For a positive whole exponent, is perfectly fine: , — a row of zeros multiplies to zero, no problem. The trouble starts elsewhere:
- ::: undefined — Step 5 needs the reciprocal, and you cannot divide by zero
- ::: ambiguous — the empty product says , but "zero to a power" pulls toward ; there is no single agreed value So is barred not because it is "bad" everywhere, but because negative and zero exponents on it break (division by zero, ambiguity). Positive-integer powers of are fine; we just cannot build the full real-exponent system on it.
Case . Here the failure is a genuine contradiction in the fraction rule. A fraction upstairs can be re-spelled (), and for a negative base the spellings disagree:
- and ::: the same exponent, reduced
- vs ::: two different answers — the law has collapsed
- the fix ::: forbid it — demand , and the ambiguity vanishes
WHY show both? So you never wander into either trap thinking "it's just like ." dies from division by zero / ambiguity; dies from multi-valued roots. Only survives every step.
PICTURE. Left panel: fine for but and flagged. Right panel: the fork landing on and with a red contradiction splash.

Step 8 — Push all the way to real (irrational) exponents
WHAT. What is , when never ends and is no fraction? We cannot "count blocks." Instead we trap between terminating decimals (which are fractions, hence block-defined by Step 6) and watch where the powers head.
WHY a limit (this tool, not another)? A limit answers precisely the question "the inputs are creeping toward a target we can't hit exactly — where is the output heading?"
Recall Why may we say "limits respect
and "? That is the content of the limit laws: if and as the rationals close in, then and too. In words: the limit of a product is the product of the limits, and likewise for sums. These are proved in Limits and Continuity; the matching statement that itself has no jumps lives in Exponential Functions and their Graphs. We use them here — we do not re-prove them.
- each entry ::: a rational exponent (a terminating decimal) — already block-defined by Step 6
- the arrow ::: "closes in on" — the limit
- ::: the single value all these approach ()
Because the limit laws let and pass through, every law survives unchanged — e.g. , exactly as if were a whole number.
PICTURE. A staircase of dots climbing the smooth curve (base , increasing) beside a second curve (base , decreasing) — both squeezing onto the vertical line .

The base (see Number e and Natural Exponential) and logarithms (see Logarithms — Definition) are built on exactly this machinery.
The one-picture summary
Everything above compressed into a single flow: blocks → join them → count → force the gaps (zero, negative, fraction) → forbid → limit to reals.

Recall Feynman retelling — the whole walkthrough in plain words
A power is a row of identical blocks: is three copies of multiplied. To multiply two powers you just shove the two rows together and count all the blocks — blocks then more is , and that "count them all" is why the little numbers add. Now we don't want holes in the rule. An empty row must count as — multiplying no factors at all (the empty product) starts at the do-nothing value — so . An anti-row must cancel a row down to nothing, so (a flip, never a negative). Repeating a row times lays blocks in one long line, so ; reading that backwards makes a fraction upstairs a root, and is just copies of the -th root. We only allow positive bases: dies on negative or zero exponents (dividing by zero, or the ambiguous ), and dies because the same fraction spelled two ways (like and ) gives two clashing answers. Finally, for an endless irrational exponent like , we trap it between terminating decimals — which are secretly fractions — and the marching values (climbing if base , descending if base ) get squeezed into a gap shrinking to zero, so they close in on one number. Since the limit laws let and pass through, all the block-counting laws survive the trip. One idea, all the way up: count the blocks.
Connections
- Laws of Exponents — Review with Real Exponents — the parent this walkthrough visualises.
- Rational Exponents and Radicals — the fraction-to-root step (Step 6).
- Exponential Functions and their Graphs — the smooth curve behind Step 8.
- Limits and Continuity — why the irrational limit lands on one value.
- Number e and Natural Exponential — the special base built on these laws.
- Logarithms — Definition — the inverse machine that undoes all of this.
- Binomial Theorem — the correct expansion of , not .