The star that was cut before it was proved

Fold a square in half, then into a wedge, then again, until the sheet is ten layers of thirty-six degrees. Make one straight cut across the wedge. Open it out and a regular five-pointed star falls into the hand.

It is a good trick and it is old. The theorem that says any drawing made of straight lines can be released by a single cut, given the right folding, is from 1998.

Why the star comes out regular #

The construction gives regularity for free, and understanding why is the key to everything else in the essay.

Folding into 2k equal wedges about a point means that whatever is cut in one wedge is reproduced in all of them, reflected alternately. So the released outline has 2k vertices and is invariant under a rotation of 360°/k — which is to say it is regular by construction, whatever the cut was.

The cut only chooses which regular star. Its angle to the fold line fixes the ratio between the outer radius and the waist, and nothing else is free.

What the fold decided and what the cut decided #

That division is worth making explicit because it inverts the intuition.

A person making this star thinks they are cutting a star shape. They are not: they are cutting a straight line. Every feature that makes the result a star — the five points, their equality, their equal spacing, the regularity of the waist — was decided by the folding, before the scissors came out.

The cut contributes one number. The generator computes the waist as a function of the cut angle and asserts the outline’s regularity independently, so the picture demonstrates the division rather than asserting it.

The waist is the only free number #

Working out how much freedom the method really has is a short calculation and it is more restrictive than it appears.

Within one wedge the cut is a straight segment joining a point on one fold line to a point on the other. Two endpoints is two numbers, but scaling the whole figure is not a shape change, so one of them is a size and only the ratio matters. The shape of the released star is therefore a one-parameter family.

That parameter is the waist: how deep the notches between the points go. Everything a person can decide when cutting a five-pointed star this way is one number, and every other feature was fixed when the last fold was made.

It also means the method cannot make a wrong star. Any cut that meets both fold lines releases a regular star of some proportion; a shaky hand changes the waist and nothing else. That robustness is why the trick is a trick — it works for people who have never done it before, which almost nothing else in this subject does.

The waist, written down #

The sine rule the generator uses is short enough to give, and giving it turns the one free number into something a person can aim at.

Fold into $2k$ wedges, so each is $180°/k$ wide, and cut across one at an angle $\beta$ to a fold line. The cut, the two fold lines and the apex make a triangle whose angles are $180°/k$, $\beta$ and the rest, so

$$\frac{r}{R} = \frac{\sin\beta}{\sin\!\big(180° - 180°/k - \beta\big)}$$

For five points the wedge is 36° and this is $\sin\beta / \sin(144° - \beta)$, which rises steadily with $\beta$ from a needle-sharp star toward no star at all. At $\beta = 72°$ the waist equals the outer radius and the released shape is a regular decagon; past that the points turn inward.

So the usable range is a cut anywhere between zero and 72°, and everything in it is a regular five-pointed star of some sharpness.

Three cuts that are already golden #

The interesting part is where the easy angles land, and they land somewhere worth knowing.

Cutting at 18° — half the wedge — gives $\sin 18°/\sin 126°$, which is exactly $(\sqrt5-1)/(\sqrt5+1) = 1/\varphi^{2} = 0.382$. That is the waist of the pentagram, the five-pointed star everybody draws, in which the points’ inner corners sit where the diagonals of a regular pentagon cross.

Cutting at 36° — the full wedge — gives exactly $1/\varphi = 0.618$. Cutting at 54° gives exactly $\varphi/2 = 0.809$, which is the fat star the figure at the top of this essay draws.

All three are exact golden-ratio values, and none of it is arranged: they follow from the sines of multiples of eighteen degrees, which is where the golden ratio lives.

So the three cut angles a folder can hit without measuring anything — half a wedge, one wedge, one and a half — give the three stars the eye recognises. The pentagram is not a difficult target; it is the cut at half the wedge, and half of a fold is the one subdivision a hand makes accurately.

That is a small addition to the trick and it changes what the free parameter is worth. The method’s one number is not merely robust, in the sense that any cut gives a regular star; it is calibrated, in the sense that the cut angles a person naturally produces are the ones that give the classical proportions.

Why this is not the theorem #

The 1998 result is a much stronger statement and the gap between them is the whole reason both dates are right.

The fold-and-cut theorem says: take any drawing made of straight line segments — a letter of the alphabet, a polygon with reflex corners, several disconnected shapes — and there exists a folding of the sheet that brings every segment of the drawing onto a single line, so one straight cut releases it exactly.

The construction is the straight skeleton of the drawing plus perpendiculars from its nodes, and it works because those creases carry every edge onto the same line. Nothing about symmetry is required and none is used.

The wedge method’s ceiling #

Set against that, the traditional method’s reach is small and precisely describable.

It releases exactly the shapes with the rotational symmetry that the folding imposes, whose boundary within one wedge is a single straight segment. That is stars, regular polygons, and a handful of related figures. A house, a letter, a shape with reflex corners in the wrong places — none of them.

So two centuries of successful practice on stars is entirely compatible with the general problem being open, because the general problem is a different problem. Compressing the two produces the claim that the theorem is ancient, which is the same error the rest of this field is about with its sign reversed.

Cutting through many layers is the real skill #

The part of the traditional method that actually takes practice is not the folding, and noticing that reframes what the trick is.

Ten layers of paper is a substantial thickness to cut cleanly with scissors, and the cut has to be straight through all of them. A blade that wanders produces points of unequal length; a blade that compresses the stack ahead of it produces a cut that is straight on the top layer and drifting on the bottom.

So the difficulty in the traditional method sits entirely in the cut, while all the geometry sits in the fold. That is an unusual division of labour and it is probably why the trick survived as a demonstration: the impressive part is easy and the fiddly part is a manual skill anybody can acquire in a few attempts.

The story, and why it is the record’s weakest entry #

The specific claim attached to this trick deserves separate treatment, because it is the only entry in this site’s record whose source is a secondary account.

The story is that a five-pointed star was cut this way in a Philadelphia upholsterer’s workshop in 1776, to settle a design question. Its earliest source is an account given in 1873 — ninety-seven years after the event — by a descendant of the family, with no surviving primary document behind it.

The record marks it as secondary, and the kind field exists for exactly this distinction. A ninety-seven-year gap between event and first telling, from an interested party, is a specific and well-understood species of weak evidence, and filing it beside a printed book with a publication date would flatten the thing that matters most about it.

Two claims wearing one sentence #

It is worth separating the components of the story properly, because they fail in different ways and get defended together.

The first component is that this technique existed in 1776. That is plausible on general grounds — cut-paper decoration was widespread, the method is easy to find by accident, and nothing about it requires anything unavailable. No specific evidence supports it and none is really needed for plausibility.

The second is that a particular person used it on a particular occasion to settle a particular design question. That is a claim about an event, it is the part that carries all the interest, and it is the part with one late secondary source behind it.

Defenders of the story tend to argue the first and treat the second as established. That is a general move worth recognising: establishing that something was possible and letting it stand in for evidence that it happened.

What is not in doubt #

Separating the parts matters, because the essay is not an argument that the trick is recent.

That the wedge method was in use long before 1998 is not seriously disputed. Houdini put the one-cut star into a book of paper magic in 1922 and treated it as known rather than as new; similar cut-paper work appears across several traditions well before that.

What is in doubt is the specific anecdote, which is a claim about one workshop on one occasion, and which the evidence does not support. Those are different claims and the second is often used to authenticate the first, which it cannot.

There is a further reason the anecdote is durable, and it has nothing to do with folding. The story does useful work for a national narrative, which is a strong preservative and a poor filter. This site’s record has no way to weigh that and does not try; it records that the source is one, secondary, and late, and lets a reader draw the conclusion.

Kirigami is not a lesser art #

There is a categorical judgement buried in how this material usually gets discussed, and it is worth challenging.

Cutting is filed as kirigami and treated as a separate and lesser thing, on the grounds that real origami does not cut. That rule is twentieth-century, generative rather than historical, and its own oldest document breaks it.

And the mathematics does not respect the boundary at all. The fold-and-cut theorem is one of the most striking results in the field and it is a result about cutting. The straight skeleton it depends on is the same object that appears in the design of bases. Treating cutting as outside the subject would remove a theorem and a construction that the subject cannot do without.

What the general construction costs #

The comparison becomes concrete once the two are priced in creases.

The wedge method uses 2k creases, all radial, all through one point. For a five-pointed star that is ten folds and a child can make them.

The general construction’s crease count grows with the complexity of the outline: every skeleton node contributes arcs and perpendiculars, and a shape with reflex corners has a skeleton with genuine structure in it. The assignment then has to be found rather than stated, because at a skeleton node the arcs and the perpendiculars alternate in equal numbers — which is a tie, and Maekawa never allows a tie.

So the traditional method is not merely a special case; it is the special case in which the construction becomes something a person can discover without knowing it exists.

What the folding is doing, in the theorem’s terms #

There is a nice way to see the traditional method as a degenerate case of the general one, and it makes both clearer.

For a regular star, the straight skeleton is almost trivial: by symmetry all its nodes collapse onto the centre. So the general construction’s creases become the radial folds through the centre plus the perpendiculars — which is exactly the wedge fold.

The traditional method is therefore the fold-and-cut construction applied to a shape whose skeleton has one node. It is not a different technique; it is the theorem’s construction in the one case a person could have found by hand.

One consequence of that reading is worth keeping. Because the wedge fold is the theorem’s construction in the symmetric case, the traditional method was never wrong or approximate — it was exactly right, on exactly the shapes it applies to, for the whole two centuries. What 1998 added was not correction but reach.

What the generator checks #

The regularity is computed rather than assumed, and the sequence is worth stating because it is what makes this a checkable historical claim.

The wedge angle is derived from the requested number of points. The cut angle is required to meet both edges of the wedge, or there is no closed shape to release. The inner radius is then computed from the cut angle by the sine rule — an output rather than a parameter. The outline’s vertices are built from those two radii, and the generator then checks independently that the radii fall into exactly two values and that the turning at every vertex of a kind is identical, both to a part in a million million.

A drawing produced from coordinates that looked right would fail those. The picture is the construction executed.

Why nobody thought there was a question #

The last historical point is about what makes a problem visible, and it is the reason the two dates are so far apart.

A technique that works does not advertise its own limits. A person who can cut a five-pointed star, a six-pointed star and a snowflake has no experience of failure to prompt the question what shapes can this reach. The failures never happen, because nobody attempts a shape without the symmetry — the method does not suggest it.

So the general problem is invisible from inside the practice. It becomes a question only when somebody asks it from outside, in the vocabulary of which drawings are reachable, which is a computational-geometry question and not a paper-cutting one.

That is the same shape as Beloch’s paper being unfindable: the obstruction was formulating the search, not performing it. In both cases the gap between practice and theorem is a gap in the question rather than in the answer.

The idealisation, named #

The folded wedge here is ten layers of zero thickness, and a real sheet folded into ten layers is about a millimetre of stack.

That matters more here than in most figures, because the cut passes through every layer at once and a real cut through a millimetre of paper is not a plane — the blade wanders, the inner layers are cut at a slightly different place from the outer ones, and the released star’s points are not quite equal. Anybody who has done this has seen it.

So the regularity asserted above is a property of the construction, and a folded instance approximates it. The interesting part is that the error is bounded by the cut’s wander rather than by any inaccuracy in the folding, which is the usual situation in this subject reversed.

Where this goes next #

This closes the folklore ladder. One straight cut is the theorem in its own terms, and the oldest book cuts the paper is the rung below, where cutting turns out to be in the tradition from its earliest surviving document.

The surprising connection to end on: this is the one entry in the record where the practice is genuinely old and the mathematics is genuinely new, and the popular error is therefore to date the theorem too early rather than the craft. Every other claim in this field runs the other way. What produces the inversion is that a trick which works is self-evidently possible, so nobody thought there was anything to prove — and the thing that took until 1998 was noticing that the general case was a question at all.