Intuition The one core idea
This whole topic is about one very special curve whose steepness at any point is exactly as big as how high the curve is at that point — the higher it climbs, the more sharply it climbs. Every symbol needed to make that sentence precise (powers, a special number, "slope", and "closing in on a value") is built from scratch below, in an order where each idea only uses ones already introduced.
This page assumes nothing . We list every symbol and idea the parent note leans on, in an order where each one only uses things already built. Read top to bottom; by the end the parent note should feel obvious.
highlight markup
Throughout this vault, wrapping a phrase in double equals like this marks it as a key term to remember — when you review, that phrase is hidden so you can test whether you can recall it. It changes nothing about the meaning; it is just a study cue.
Definition Exponent notation
a b (whole-number power)
When b is a whole number , a b means "start at 1 and multiply by a , a total of b times." The number a sitting on the ground is the base ; the small number floating up top, b , is the exponent (also called the power or index ).
The picture: think of a ladder. Each rung multiplies you by a . a 3 means climb 3 rungs from the floor:
a 3 = 1 × a × a × a .
Intuition Why the exponent lives
up top
The position matters enormously. In x 2 the base moves (a growing square). In 2 x the exponent moves (a doubling process). These behave completely differently — the whole parent topic is about the second kind, where the variable is the height of the ladder, not the size of a step.
a x to all real x (needs a > 0 )
"Multiply a by itself b times" only makes sense when b is a whole number — you cannot multiply "half a time." We stretch the meaning in stages:
Fractions: a 1/2 = a (the number that, squared, gives a ); a 1/ n is the n -th root.
Real exponents: for an irrational power like a 1.4142 … we take fractions closing in on it and see what value a those fractions homes in on.
This stretching only works when the base is positive , a > 0 . If a were negative, roots like a 1/2 = a would not be real numbers, and the graph of a x would tear into gaps. So throughout this whole topic we require a > 0 (and a = 1 for a genuinely growing/shrinking curve).
Intuition Why the "closing-in" for real exponents actually lands on a value
As you feed a x a ladder of fractions creeping toward an irrational number — say 1.4 , 1.41 , 1.414 , … toward 2 — the results a 1.4 , a 1.41 , a 1.414 , … do not scatter about; they get squeezed into an ever-tighter band and settle on one value. That happens because for a > 0 the powers rise (or fall) smoothly with no jumps, so nearby exponents give nearby answers. This is what lets us say a x has a single, well-defined value for every real x , with no holes in the graph — a fact we quietly rely on the moment we plug a tiny real step h into a h .
a > 0 " fine print matters later
In section 5 we form the quantity h a h − 1 for very small steps h , which feeds a tiny fractional powers a h . That expression is only meaningful if a h is a real number for every small h — which is exactly what a > 0 guarantees.
Why the topic needs these:
a m + n = a m ⋅ a n is the engine of the whole derivative derivation. When the parent writes a x + h = a x ⋅ a h , that is this law with m = x , n = h .
a − n = 1/ a n explains why (as we'll see) e − x is a small positive number, never negative — a reciprocal of a positive is still positive.
a 0 = 1 explains why every exponential graph passes through the point ( 0 , 1 ) .
Worked example Warming up the laws with a friendly base
a = 2
2 x + h = 2 x ⋅ 2 h — split a combined exponent into a product. Why? The + in the exponent becomes a × outside.
2 − 1 = 2 1 1 = 2 1 = 0.5 — a negative exponent flips it into a fraction.
2 0 = 1 — zero copies of anything multiplied leaves you at the starting 1 .
4 1/2 = 4 = 2 — a fractional exponent is a root (this only makes sense because 4 > 0 ).
(We deliberately used the plain base a = 2 here, because the special base e — the main character of this topic — isn't defined until section 2.)
Definition Euler's number
e
e ≈ 2.71828 … is a fixed number, like π . It is irrational (its decimals never end or repeat), so e = 2.7 is only an approximation. It is not chosen for looking neat — it is chosen for a special growth property we unpack in section 5. Note e > 0 , so everything from section 1 applies to e x .
The picture: on the number line, e sits just past 2.7 , between 2 and 3 — and section 1 tells us e x therefore has a full unbroken graph.
Figure s01 — a technical-drawing number line with the amber marker for e ≈ 2.718 ; alongside it, the three ladders 2 x , e x , 3 x compared at x = 1 , showing e is the base wedged between 2 and 3 that will turn out to be "just right".
Now that e is defined, we can safely reuse the section-1 laws with base e :
e x + h = e x ⋅ e h (product law), e − 1 = e 1 ≈ 0.368 , e 0 = 1 .
Intuition Why a whole letter for one number?
Because among all the ladders 2 x , 3 x , 2. 5 x , … exactly one base makes the function perfectly self-copying under differentiation (section 5). That special base got its own name: e . See Euler's number e — definition & limit for its full definition.
Definition Function notation
f ( x )
A function is a machine: feed in a number x , get out one number f ( x ) . The name f is the machine; the x in brackets is what you fed it. So f ( x ) = e x says: "the machine named f takes x and returns e raised to that power."
The picture: a box with an input arrow (x ) and an output arrow (f ( x ) ).
Definition The graph of a function
The graph plots the input on the horizontal ==x -axis and the output on the vertical y -axis==, so y = f ( x ) . Each point on the curve is a pair ( x , f ( x )) — "where I stood" and "how high the machine sent me." From now on the symbol y means exactly this output height.
Intuition Two words for the same height
"y ", "f ( x ) ", and "the height of the curve" all mean the same thing. When the parent says slope equals height , "height" is just y = e x .
Definition Slope (gradient)
The slope of a straight line is run rise — how much you go up (y ) for each step you take across (x ).
slope = change in x change in y = Δ x Δ y
The symbol Δ (Greek capital "delta") just means "the change in".
The picture: a right triangle riding on the line — horizontal leg is the run, vertical leg is the rise.
Figure s02 — a cyan straight line on blueprint grid; an amber right triangle underneath shows the horizontal "run" (change in x ) and vertical "rise" (change in y ) whose ratio is the slope.
A big slope = steep climb.
A slope of 0 = flat.
A negative slope = going downhill.
curve ? Zoom in.
A curve bends, so it has no single slope. But if you zoom into any point far enough, the curve looks straight — the slope of that tiny straight piece is the slope at that point . The line touching the curve there is the tangent line , and its slope is the curve's slope at that point.
Figure s03 — the cyan curve y = e x with an amber tangent line just touching it at one white point; the tangent's slope is the curve's slope exactly there.
Intuition The topic's headline, now stated with every word earned
For y = e x : at the point where the curve's height is y , the tangent line's slope is exactly y . Height and steepness are the same number — every symbol in that sentence (y , height, slope, tangent) is now defined. That is the whole show.
We know slope for a straight line. To get the slope of a curve at a point we sneak up on it.
Definition Average slope between two points
Take the point at x and a nearby point a little distance h away, at x + h . Their heights are f ( x ) and f ( x + h ) . The straight line between them (a secant ) has slope
h f ( x + h ) − f ( x ) .
Here h is a small horizontal step; the top is the rise, h is the run.
The picture: a secant line cutting the curve at two points. As we slide the second point closer (shrink h ), the secant swings toward the tangent.
Figure s04 — two grey secant lines through y = e x for a large and a smaller step h ; as h shrinks toward 0 the secants swing onto the amber tangent line. This is the picture of a limit.
lim h → 0 symbol
h → 0 lim ( expression ) asks: "what value does this expression home in on as h gets closer and closer to 0 ?" We cannot just set h = 0 (that would put 0 on the bottom, which is illegal), so instead we watch the trend .
Definition The derivative
d x d f ( x )
The derivative is the slope of the tangent — the average slope with the gap h squeezed to nothing:
d x d f ( x ) = lim h → 0 h f ( x + h ) − f ( x ) .
The notation d x d literally reads "the rate of change with respect to x ." d x d y means the same thing.
needs limits and derivatives
"Slope equals height" is a statement about the tangent line's steepness. The only honest way to talk about the slope of a bending curve at a single point is the limit above — there is no shortcut allowed before you build the rule. This is exactly the "first principles" the parent starts from.
Feed f ( x ) = a x into the derivative and use index law a x + h = a x a h (valid since a > 0 , and valid for the real step h thanks to section 1):
d x d a x = lim h → 0 h a x + h − a x = lim h → 0 h a x a h − a x = a x call this k ( a ) h → 0 lim h a h − 1 .
Intuition Why we're allowed to pull
a x outside the limit
The limit variable is h — that's the only thing changing as we squeeze the gap to 0 . But a x contains no h at all : for a fixed input x it is just a frozen number (like a 2 or a 5 ). A limit lets any factor that doesn't depend on the limit variable slide out front unchanged, exactly like lim h → 0 ( 7 ⋅ g ( h )) = 7 ⋅ lim h → 0 g ( h ) . Here the "7 " is a x .
What just happened, in words: the derivative of any exponential is the same function back , multiplied by a constant k ( a ) that depends only on the base. The number e is defined as the base making that constant equal to 1 :
lim h → 0 h e h − 1 = 1 ⟹ d x d e x = e x .
Common mistake "I can just set
h = 0 in h a h − 1 ."
Why it feels right: we do want h tiny, so why not 0 ?
The fix: at h = 0 you get 0 0 — undefined. The limit dodges this by asking what the trend is as h shrinks, never actually landing on 0 .
A curve is concave up if it bends like a smile (a valley) — as you move right, the slope keeps increasing . Since for e x the slope equals the height, and the height only rises, the slope only rises too: e x smiles everywhere.
Below is a dependency map : read it top-to-bottom, following the arrows. Each box is one foundation from this page; each arrow means "you need the box at the tail before the box at the head makes sense."
Start top-left at "Powers a^x and index laws" (section 1). It splits two ways: it fixes the special base e (you can't pin down e without powers), and it powers function notation .
Follow the chain function → graph → slope → tangent line : each idea is drawn on the one before it — a graph needs a function, slope needs a graph, a tangent needs slope.
Tangent line feeds the limit (we slide a secant toward the tangent), and both tangent line and limit flow directly into the derivative — the derivative is the tangent's slope computed by a limit. Notice powers also arrows straight into the derivative, because the derivation reuses the index law.
Finally e and derivative meet at "Natural exponential e^x" , which delivers the headline "slope equals height" , from which concave up follows.
Powers a^x and index laws
The number e approx 2.718
What does the exponent in a b tell you to do when b is a whole number? Start at 1 and multiply by the base a , exactly b times.
Why must the base satisfy a > 0 for a x to work for every real x ? Otherwise roots like
a 1/2 = a aren't real, and the graph tears into gaps; positivity keeps
a h real for all small
h .
Why does a x have a single well-defined value even for irrational x ? Fractions creeping toward x give powers that are squeezed into one value (no jumps for a > 0 ), so the limit exists and is unique.
Rewrite a x + h using an index law, and note which exponents it's valid for. a x + h = a x ⋅ a h ; valid for any real exponents, including a tiny real step h .
Why is e − x always positive? e − x = 1/ e x , a positive divided by a positive, so it stays positive (just small).
What is a 0 for any a > 0 , and which graph point does it fix? a 0 = 1 , so every exponential passes through ( 0 , 1 ) .
Approximate value of e , and is it exact? e ≈ 2.71828 … ; it is irrational, so no finite decimal is exact.
What does f ( x ) mean? The output of the function machine f when fed the input x .
What do y , f ( x ) and "height of the curve" have in common? They are three names for the same thing — the output plotted on the vertical axis.
Define slope of a straight line in words. Rise over run — the change in y divided by the change in x .
What is the tangent line at a point on a curve? The straight line just touching the curve there; its slope is the curve's slope at that point.
What is lim h → 0 ( … ) asking? The value the expression homes in on as h gets arbitrarily close to 0 (without being 0 ).
Why may a x be pulled outside lim h → 0 ? For a fixed x it contains no h , so it's a constant with respect to the limit variable and slides out front.
Write the first-principles definition of the derivative. d x d f ( x ) = lim h → 0 h f ( x + h ) − f ( x )
What single property defines the base e ? The base whose derivative constant is 1 : lim h → 0 h e h − 1 = 1 , so d x d e x = e x .
What does "concave up" look like? A smile / valley shape — the slope keeps increasing as you move right.