How Straight Lines Make a Circle

Interactive Lesson · Mathematics

How Straight Lines Make a Circle

A lesson for students who know complex numbers: the Taylor series of e builds rotation out of perpendicular steps — and the circle shows up at the end, not the beginning.

John Rector 7 min read Interactive — runs in this page, nothing to install
01

The circle is usually an ingredient

Every trigonometry course I have ever seen starts the same way: draw a unit circle, put a point on it, and define cosine and sine as its shadow on the two axes. Rotation is assumed, and the functions are read off it. That order works, but it hides something. It makes the circle a prerequisite — a thing you must already believe in before the mathematics starts. Students memorize it. Few of them ever see what causes rotation.

This lesson runs the causality the other way. We will start with nothing but a rule for making straight arrows — each one perpendicular to the last, each one shorter — and we will watch a circle assemble itself out of them. No circle is defined anywhere in the setup. It arrives as a conclusion.

The circle isn’t an ingredient. It’s the result.

02

One rule: turn left, shrink, repeat

Take the exponential series and feed it an imaginary input. Nothing else is needed:

e  =  1  +   −  θ2/2!  −  3/3!  +  θ4/4!  +  5/5!  −  ⋯

Look at how each term relates to the one before it. To get term n+1 from term n, you multiply by exactly one thing:

tn+1  =  tn  ×  iθ/(n+1)

Multiplying by i is a pure 90° left turn. Multiplying by θ/(n+1) is a shrink. That is the entire engine: turn left, shrink, repeat. Because every factor of i lands on the cycle +1, +i, −1, −i, every single term is either purely real or purely imaginary — so when you draw the terms as arrows laid tip-to-tail, every arrow is exactly horizontal or exactly vertical, and each is perpendicular to the one before it.

Now the punchline hiding in plain sight: the horizontal steps, added up, are the series for cos θ. The vertical steps, added up, are the series for sin θ. Cosine is not a shadow of a circle here. Cosine is how far east the walk ends up; sine is how far north.

Figure 01 The arrow chain at θ = 1: six perpendicular steps, and the tip lands on a circle nobody drew
  • +1
  • +iθ
  • −θ²/2!
  • −iθ³/3!
  • +θ⁴/4! …
tip → (cos 1, sin 1) east = cos θ north = sin θ
Pure CSS, drawn from computed values. At θ = 1 radian the six steps are +1, +i, −0.5, −0.1667i, +0.0417, +0.0083i; the running total after six terms is (0.5417, 0.8416), against a true (cos 1, sin 1) = (0.5403, 0.8415). The dashed arc is added here only for comparison — the construction never references it.
03

Make the circle appear

The demo below is the whole argument, live. The left pane is the top-down view: the arrow chain for the current angle θ, one term at a time. The right pane is the side view: the graph of the partial sums of cos x. Add terms with +1 term. Then press ▶ sweep θ together with trace tip and watch the pink tip draw a curve nobody asked for. Only after it closes should you press reveal unit circle.

e =

θ 1.000 terms 2

Top-down: the arrow chain

Side view: partial sums of cos x

Σ east = · cos θ = Σ north = · sin θ = tip distance from 0 = (circle ⇒ 1)
Acid arrows are the real (east) steps that sum to cos θ; cyan arrows are the imaginary (north) steps that sum to sin θ; the pink dot is the running total, and the violet trail is where the total has been. Every joint in the chain is a right angle — small squares mark them where there is room.

One number in the readout deserves attention: the tip’s distance from the origin. The construction says nothing about that distance staying at 1. It just does — for every θ, once enough terms are in. Constant distance from a fixed point is the definition of a circle, and here it is an output of turn-left-shrink-repeat, not an axiom.

04

Why new terms never wreck old ones

The worry every student has about an infinite series: if I keep adding things forever, why doesn’t the sum wander off? Why doesn’t term nine ruin what terms one through eight built? The answer is the factorial. The n-th arrow has length θn/n!, and n! grows so much faster than θn that each correction is smaller than the last — brutally so.

Figure 02 Length of the n-th arrow at θ = 1: the factorial collapse
  • 1.000
  • 1.000
  • 0.500
  • 0.167
  • 0.042
  • 0.008
  • 0.001
  • n=0
  • n=1
  • n=2
  • n=3
  • n=4
  • n=5
  • n=6
Computed values of θn/n! at θ = 1. By the seventh arrow the step is a thousandth of the first. Closer to zero it is far more extreme: at θ = 0.2 the fifth arrow is already 0.0000667 long — it would be invisible at this scale.

There is a second, structural reason the early picture survives, and it matters more than the shrinking: adding a term never edits the terms before it. The five-term polynomial contains the three-term polynomial verbatim — same coefficients, untouched. New terms are appended, not blended in. Near zero, where θn is microscopic, the new term contributes essentially nothing, so every partial sum agrees there; farther out, where the old polynomial was starting to fail, the new term does its work. Press zoom near 0 in the demo and look: every curve lies on top of every other curve near the center. Higher-order terms are polite. They fix the frontier and leave the settled territory alone.

Figure 03 Correct digits of cos 1, by number of terms
  • 1 term (just 1)0.3
  • 2 terms1.4
  • 3 terms2.9
  • 4 terms4.6
  • 5 terms6.6
Digits of accuracy = −log10 of the error, computed exactly: errors are 4.6×10−1, 4.0×10−2, 1.4×10−3, 2.5×10−5, 2.7×10−7. At θ = 1, each extra term buys roughly two more correct digits.
05

The classroom arc

The sequence matters. The whole point is that students see the circle arrive, so do not show it early. This is the order I would run:

  1. Start with one arrow

    Set terms to 1. One acid arrow of length 1 pointing east. Ask the room: where is the circle? There isn’t one. There is nothing here but a straight line.

  2. Add terms one at a time at a fixed angle

    θ = 1 works well. Each press of +1 term adds one perpendicular correction, visibly smaller than the last. Say the rule out loud each time: turn left, shrink.

  3. Sweep θ with the trace on

    The pink tip starts drawing. Do not name the shape. Let someone in the room say the word “circle” first — it will not take long.

  4. Only now reveal the unit circle

    The dashed circle lands exactly on the trace. The thing they were told to assume in trigonometry has just been manufactured from straight lines in front of them.

  5. Zoom near zero on the graph

    Every partial sum agrees near the center. This is the quiet lesson underneath: each term has a job, the jobs do not overlap, and the low-order terms are load-bearing and never touched again.

Where this picture runs out

Honesty for the front of the room. First: the demo caps at 14 terms and |θ| ≤ π, and that flatters the series. At θ = 2π the partial sums spiral roughly 85 units away from the origin before winding back to the circle — convergence is guaranteed, but it is not always tidy, and a student who tries a big angle at home deserves to know that is a feature of the mathematics, not a bug in the page.

Second: seeing is not proving. The demo shows that the factorial wins and the tip lands at distance 1; it does not show why. Those proofs — the ratio test for convergence, and the fact that |e| = 1 — are where this lesson should send students next, and they will arrive already believing the conclusion, which is the right order to learn a proof in.

Euler wrote this series down in the 1740s. Nearly three centuries later it is still, I think, the shortest honest answer to a question most students never get to ask: what actually causes rotation? Not a circle. A rule. Turn left, shrink, repeat — and the circle takes care of itself.

Author: John Rector

John Rector is a Charleston-based entrepreneur, author, and AI strategist. He co-founded E2open, the supply-chain software company acquired for $2.1 billion in 2025, and in 2026 opened Charleston AI, a 3,000-square-foot lab that helps people and organizations understand and use artificial intelligence. He is the creator of The Reality Equation — a lecture series, book, and curriculum exploring attention, prediction, and how reality is experienced — and the author of more than two dozen books. He writes and speaks widely on artificial intelligence, attention, and the future of human work.

2 thoughts on “How Straight Lines Make a Circle

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Discover more from John Rector

Subscribe now to keep reading and get access to the full archive.

Continue reading