Intuition The ONE core idea
A derivative is the steepness of a curve at a single point , found by measuring the slope between two points and then sliding them together until they merge. Everything below is the vocabulary and pictures you need so that when the parent note writes lim h → 0 h f ( x + h ) − f ( x ) , not one symbol is a stranger.
This page assumes you know nothing beyond "a graph is a picture of numbers." We will earn every letter, one at a time, and each new idea leans on the one before it.
f means
f is a machine : you feed it a number, it gives back one number. We write f ( x ) to say "the number that comes out when x goes in." The x inside the brackets is just the input slot .
f = the machine's name.
x = whatever number you drop in.
f ( x ) = the number that comes out (the output ).
Think of a vending machine. Press button x = 3 , out comes a snack f ( 3 ) = 9 (if the machine is f ( x ) = x 2 ). The rule is fixed; only your input changes.
Why the topic needs it: the derivative is a fact about a function . No machine, nothing to differentiate.
f
For every input x we get an output f ( x ) . Plot the point ( x , f ( x ) ) : go across by x , then up by f ( x ) . Do this for all x and the dots join into a curve . That curve is the graph.
The horizontal axis measures the input x (how far across).
The vertical axis measures the output f ( x ) (how far up).
A single point on the curve is written ( x , f ( x ) ) — an "address" (across, up).
Intuition Why a curve, not a line?
If bigger inputs give proportionally bigger outputs, you get a straight line. But f ( x ) = x 2 grows faster and faster , so the picture bends upward — a curve. Different places on that curve tilt by different amounts, and that tilt is the whole story of this topic .
Why the topic needs it: "slope" and "steepness" are things you see on the graph. Without the picture there is nothing to measure the steepness of.
Definition Slope of a straight line
Slope answers: "for every step I take sideways, how much do I climb?" Pick two points on the line. The run is how far right you moved; the rise is how far up you climbed. Then
slope = run rise = change in sideways change in height
Run = change in the input (horizontal).
Rise = change in the output (vertical).
A big slope = steep. A slope of 0 = flat. A negative slope = going downhill as you move right.
Intuition Why "over"? Why divide?
Dividing rise by run gives climb per single sideways step . That makes slopes comparable: a staircase that climbs 2 m over 4 m (2 1 ) is gentler than one climbing 3 m over 1 m (3 ). Dividing is what turns "total climb" into "steepness."
Common mistake Slope needs TWO points
You cannot compute rise-or-run from a single dot — a change needs a "before" and an "after." This single fact is the entire reason the topic exists: a curve's slope changes at every point, yet slope seems to need two points. The trick (Section 6) resolves this.
Why the topic needs it: the derivative is a slope. Every worked example in the parent ends with a slope value.
See Average vs Instantaneous Rate of Change — a two-point slope is an average rate; the derivative is the instantaneous one.
h stands for
h is just a small distance we step to the right along the input axis. Starting at input x , the nearby input is x + h . Nothing mysterious — h is a number like 0.1 or 0.001 that we will eventually shrink.
x = your chosen point (stays put).
x + h = the buddy point a little to the right.
h = the gap between them (the run , in the language of Section 3).
Intuition Picture the buddy
Imagine your finger on the curve at x . Now a second finger lands a tiny bit to the right at x + h . As h gets smaller, the two fingers crawl toward each other.
h is not zero — yet
h is small but alive (non-zero) while we do algebra. Only at the very last step do we ask what happens as it heads to zero . Setting h = 0 too early breaks everything (Section 7).
Why the topic needs it: h is the controllable "distance between the two points" that we will squeeze to make the two points merge.
f ( x + h ) means
Feed the buddy input x + h into the machine. The result f ( x + h ) is the height of the curve at the buddy point . You get it by replacing every x in the rule with the block ( x + h ) .
Worked example Concrete substitution
If f ( x ) = x 2 , then f ( x + h ) = ( x + h ) 2 . Expanding: ( x + h ) 2 = x 2 + 2 x h + h 2 .
If f ( x ) = x 1 , then f ( x + h ) = x + h 1 .
Common mistake Substitute EVERYWHERE
For f ( x ) = x 2 , the buddy output is ( x + h ) 2 , not x 2 + h . Treat x + h as one sealed block and expand it honestly.
Why the topic needs it: to measure a slope you need the height at both points — f ( x ) at yours and f ( x + h ) at the buddy's.
Now we combine Sections 3, 4, 5. The two points are:
Point A: address ( x , f ( x ) ) .
Point B: address ( x + h , f ( x + h ) ) .
The rise (climb from A to B) is f ( x + h ) − f ( x ) . The run (sideways from A to B) is ( x + h ) − x = h . So the slope of the straight line through A and B is:
A secant is any straight line that cuts through a curve at two points. Its slope is the average steepness of the curve between those two points.
Why the topic needs it: this is the object the whole topic is built on. The derivative is what this quantity approaches as the buddy slides in. See Secant and Tangent lines .
lim h → 0 ( ⋯ ) asks
Read it aloud: "the value the expression settles on as h gets closer and closer to 0 — but never actually is 0 ." The arrow → means "approaches." We care about the destination , not the endpoint.
Intuition Why we need a limit here
As the buddy point slides toward your point (h → 0 ), the secant line rotates and, in the merged position, just touches the curve — that touching line is the tangent . But we can't land exactly on h = 0 (Section 8), so we ask where the slope is heading . The limit is the tool that answers "heading toward what?" — exactly the question this topic needs and no other tool answers it. See Limits — formal definition and one-sided limits .
The tangent touches the curve at one point and matches its steepness there. Its slope is the derivative. As h → 0 , secant → tangent.
Why the topic needs it: without a limit, "slope at a single point" is impossible — the run would be 0 . The limit is the bridge from two-point slope to one-point slope.
If we set h = 0 before simplifying, the run becomes 0 and the rise becomes f ( x ) − f ( x ) = 0 , giving 0 0 .
0 0 is forbidden
Dividing by 0 has no answer, and 0 0 could "be anything" depending on how you approach it. It is called an indeterminate form — undecided until you do more work. See Indeterminate forms 0 over 0 .
Intuition The escape route
First do algebra to cancel the troublesome h (factor it out of the top, cancel with the bottom). Once cancelled, the expression is safe and you can let h → 0 peacefully. That cancellation is the beating heart of every worked example in the parent note.
Why the topic needs it: it explains the strict order — simplify first, take the limit second — that makes the whole method work.
f ′ ( x )
The little mark ′ (a prime ) means "the derivative of." So f ′ ( x ) is read "f -prime of x " and means the slope of f 's curve at the point x . It is itself a new machine: input a point, output the steepness there.
f ′ is NOT a smaller copy of f
f ( 10 ) = 100 but f ′ ( 10 ) = 20 for f ( x ) = x 2 — one is a height , the other a steepness . Never read f ′ as "a shrunken f "; read it as "how fast f is changing." A curve being differentiable everywhere it's smooth ties to Continuity and Differentiability .
Why the topic needs it: f ′ ( x ) is the answer the topic produces — the derivative as a function of position.
Now the parent note's headline makes sense with no strangers left:
f ′ ( x ) = lim h → 0 h f ( x + h ) − f ( x )
Read it as a sentence: "The steepness of f at x (f ′ ( x ) ) equals the value that the secant slope (the fraction ) heads toward (lim ) as the buddy point slides in (h → 0 )."
Graph turns machine into a curve
h is a tiny step to x plus h
f of x plus h is height at buddy
Difference quotient secant slope
Zero over zero is forbidden
Test yourself — cover the right side and answer out loud.
What does f ( x ) mean in plain words? The number a machine f gives back when you feed it the input x .
On a graph, what does the point ( x , f ( x ) ) tell you to do? Go across by x , then up by f ( x ) — that dot is on the curve.
Define slope without symbols. Rise over run — how much you climb for each sideways step.
What is h , and is it zero? A tiny sideways distance to the buddy point x + h ; it is small but non-zero while we do algebra.
How do you compute f ( x + h ) for f ( x ) = x 2 ? Replace every x with ( x + h ) : ( x + h ) 2 = x 2 + 2 x h + h 2 .
What geometric object does h f ( x + h ) − f ( x ) measure? The slope of the secant line through the two curve points.
What does lim h → 0 ask? What value the expression settles on as h gets arbitrarily close to 0 (never equal to 0 ).
Why is setting h = 0 immediately forbidden? The run becomes 0 , giving the indeterminate form 0 0 .
What does f ′ ( x ) represent? The slope of f 's curve at the point x — its steepness, not its height.
As h → 0 , the secant line becomes the ___ line. tangent