[Beginner → Intermediate] — What calculus is actually about, before any of the machinery. Two questions, one idea underneath both.

Before you start

  • You can manipulate algebraic expressions — algebraic manipulation.
  • You can read a graph and find the gradient of a straight line.
  • You know what a function is — sets.

What you will be able to do after this lesson:

  1. State the two central problems of calculus and say why each is hard.
  2. Explain why the gradient of a curve is not defined the way a line’s is.
  3. Explain what a limit is informally, and why both problems need one.
  4. Say what the Fundamental Theorem of Calculus claims, before being able to prove it.

1. Two questions that look unrelated

The tangent problem. How steep is a curve at a single point?

For a straight line this is easy: pick two points, divide the rise by the run. The answer is the same wherever you pick them.

For a curve it is not easy, and the reason is worth sitting with. Steepness is a property of two points — it is a ratio of change. A single point has no change. Asking for the gradient “at a point” is asking for , which is not a number.

   y                      the gradient at P:
   |        ,-'              rise / run  needs TWO points
   |      ,'                 but we want ONE
   |    ,'  P
   |  ,'
   |,'
   +------------ x

The area problem. What is the area under a curve?

For a rectangle, area is base times height. For a curve, the “height” changes continuously, so there is no single height to multiply by.

Both problems fail for the same reason: the quantity you want to divide by, or multiply by, is not constant. Calculus is the machinery for handling exactly that.


2. The idea underneath both: approximate, then refine

The tangent problem. You cannot use one point, so use two — one at and one nearby. That gives a secant line, whose gradient is close but wrong. Then slide the second point towards and watch what the gradient approaches.

   secant through P and a nearby point:   gradient is close
   slide the point closer:                gradient is closer
   ...                                    approaches a value

The area problem. You cannot use one rectangle, so use many thin ones. Their total is close but wrong. Then make them thinner and watch what the total approaches.

In both cases the answer is not computed directly. It is approached. That approaching is called a limit, and it is the single idea calculus is built on. Everything else — derivatives, integrals, all the rules — is a consequence.

NOTE

This is why limits come first and feel unmotivated. They are the foundation, so they are taught before the buildings that stand on them.

If limits feel like pointless formalism, come back to this page. A limit is how you answer a question whose direct calculation gives or an infinite sum — and both of calculus’s central questions do exactly that.


3. The two halves, named

QuestionAnswerNotation
Differential calculusHow fast is this changing right now?the derivative,
Integral calculusHow much has accumulated in total?the integral

Differentiation takes a quantity and gives its rate of change. Position → velocity. Velocity → acceleration.

Integration takes a rate and gives the accumulated total. Velocity → distance travelled. Flow rate → volume delivered.

4. The result that ties them together

Those two operations look like different subjects. They are not.

The Fundamental Theorem of Calculus: differentiation and integration are inverse operations. Accumulate a rate of change and you recover the original quantity; take the rate of change of an accumulation and you recover the rate.

Read it in words: if you accumulate from up to , and then ask how fast that accumulation is growing at , the answer is itself — because is precisely the rate at which you are currently accumulating.

This is the central result of the subject, and it is why finding areas — which sounds geometric and hard — reduces to reversing differentiation, which is largely mechanical.


5. Where you already rely on it

  • Physics. Velocity is the derivative of position; acceleration of velocity. Every equation of motion is calculus.
  • Machine learning. Training a model means minimising a loss function, and you minimise by following its gradient downhill. Backpropagation is the chain rule applied systematically — see ai-ml.
  • Signal processing. Fourier analysis is built on integrals — see DSP.
  • Probability. A continuous distribution’s probability is the area under its density curve; the total is an integral equal to 1.
  • Anything with “rate” or “total” in it. Interest, growth, decay, flow, dosage, throughput.

6. What this course does and does not assume

Assumes: algebra, functions, graphs.

Does not assume: any prior calculus.

A warning worth taking seriously: calculus is where mathematics stops being a set of procedures to apply and starts being a structure to understand. You can pass exams by memorising rules. You cannot use it that way. Every rule in this course is derived, and the derivations are the content.

Before moving on

  • State the tangent problem and say why a single point has no gradient.
  • State the area problem and say why one rectangle is not enough.
  • Explain what “approach a limit” means informally.
  • State what the Fundamental Theorem of Calculus claims.

Recap: Calculus answers two questions that classical algebra cannot: the exact steepness of a curve at a point, and the exact area under one. Both fail directly — one gives , the other an infinite sum — and both are solved the same way: approximate with something you can compute, then refine and see what the answer approaches. That approaching is a limit, and it is the foundation of everything that follows. Differentiation and integration turn out to be inverses of each other, which is the Fundamental Theorem.

Next: The Limit of a Function — the idea both halves depend on, made precise.

↑ Back to Mathematics