Origami & Folding
what one sheet can do
Who found it, and when

The kindergarten was a geometry class

Froebel put paper folding into mass education in the 1830s, and did it as mathematics rather than as craft. His three categories — the folds of life, of beauty, and of knowledge — are the first systematic treatment of folding anybody wrote down, and the third one is a geometry syllabus.

Assumes Nothing here is as old as it sounds.

18 min read 6 figures Folding beats the compassOne sheet, no cuts

The European half of this subject does not begin with an art. It begins with a curriculum.

Friedrich Froebel, inventing the kindergarten in the 1830s, needed activities that would put mathematical structure into the hands of very young children before those children could read. Paper folding was one of them, and it entered mass education as a geometry lesson — not as decoration, not as craft, and not as an import from anywhere.

Dividing a square into 4, exactlyHaga's theorem. Folding one corner onto a point part-way along an opposite edge produces exact rational divisions elsewhere on the sheet — so a square can be divided into any whole number of parts by folding alone, with no measurement and no accumulated error.4 equal partsestimated by eyethe left-hand divisions are exact — a consequence of the fold, not of carethe right-hand ones are a guess, and the error compoundsvalleymountain
Fig. 1 The sort of thing the syllabus was for. A square, and the points a sequence of folds puts on it — halves, quarters, eighths — each of them exact, none of them measured, and every one of them available to a child with no ruler and no arithmetic.

Three kinds of fold

Froebel’s treatment divides folding into three categories, and the division is sharper than it sounds.

The folds of life produce objects: a boat, a box, a hat. They are the folds a child recognises as making something.

The folds of beauty produce pattern: symmetric arrangements, repeated forms, the results of folding and unfolding to leave a design of creases. These are the ones that look most like modern tessellation work.

The folds of knowledge produce geometry: a square divided into halves and quarters, a right angle bisected, a triangle exhibited as half a square, congruence demonstrated by superposition. These are the ones the syllabus was actually for, and they are a construction sequence.

Not an import, and not an invention

One thing the kindergarten story is regularly made to do, and cannot.

It is sometimes offered as evidence that European paper folding is independent of the Japanese tradition, and sometimes as evidence that it is derived from it. The material supports neither. Froebel assembled a syllabus from folds that were already about in German-speaking Europe — the Fröbelstern, the boat, the box — and the question of where those came from is not answered by anything he wrote, because he was not asking it.

What the record supports is narrower and is enough: from the 1830s there is a systematic, documented, institutionally transmitted body of paper folding in Europe. Before that there are scattered objects. The syllabus is the first thing anybody can point at that is a body rather than a handful.

Why folding rather than drawing

The pedagogical argument is the interesting part, and it is one this site has independently arrived at from the other direction.

A construction made by drawing requires an instrument, a steady hand, and the acceptance that the drawn line is approximately the intended one. A construction made by folding requires none of those. Bringing an edge onto an edge is the bisection; the crease is where it is because the paper was brought into coincidence, not because somebody aimed.

So a folded construction is exact in a way a drawn one is not, and it is exact for a five-year-old. That is the same observation this site makes about Haga’s construction and about dividing a strip without measuring, arrived at a hundred and fifty years later.

Folding a strip into thirdsA guess, and then halving. Each fold moves the mark to the midpoint of one of the two pieces, and each fold halves the distance to the exact division — so the error falls geometrically from whatever the first guess was. The sequence of halvings is read off the binary expansion of the fraction rather than chosen, and the halving of the error is asserted rather than observed.guessoff by 0.16667fold 1off by 0.08333fold 2off by 0.04167fold 3off by 0.02083fold 4off by 0.01042fold 5off by 0.00521fold 6off by 0.00260solid mark — the fold is aiming at 1/3; dashed — at where 1/3 has gonehalving word R L — period 2, from 2^2 − 1 = 3 × 1every fold halves the error exactly, so 6 folds divide it by 64
Fig. 2 Why folding rather than drawing: thirds, which a ruler gives approximately and this gives to any tolerance asked. Six folds from a guess, each halving what is left of the mistake.

A folding sequence is a proof

There is a stronger claim available and Froebel’s material comes close to making it.

A sequence of folds that produces a stated configuration is a constructive proof of that configuration’s existence, and it is one whose steps can be checked by performing them. That is a rare property. Most proofs available to a child are not proofs at all but demonstrations, and most demonstrations rely on the accuracy of a picture.

Here the accuracy is structural. If bringing corner to corner produces the crease, then the crease bisects, and there is no measurement anywhere in the chain that could be slightly wrong.

What a five-year-old can be shown

The choice of folding over drawing is not only about exactness, and the second reason is the one that makes the material work.

A folded construction is reversible and inspectable. The child unfolds the paper and the whole history of the construction is there as creases: which fold came first, what was brought onto what, where the resulting points fell. A drawn construction leaves the same marks whether they were made in the right order or not, and a wrong one looks exactly like a right one.

So the paper keeps a record of its own derivation. That is a remarkable property for a teaching medium and it is the same property that makes the crease pattern the canonical object of the modern subject — a pattern is a construction’s transcript, complete, and readable by anybody who did not watch it being made.

The categories are not arbitrary

Froebel’s three-way division looks like a piece of nineteenth-century tidiness and turns out to cut along a real seam.

The folds of life end in a three-dimensional object and are judged by whether the object is right. The folds of beauty end in a flat pattern of creases and are judged by symmetry. The folds of knowledge end in a point or a line on the sheet and are judged by exactness.

Those are three different success conditions, and they map with surprising directness onto three things this site keeps separate: the folded state, the crease pattern, and the construction. An essay here about a folded state and one about a construction are doing different work with different standards of evidence, and the division was drawn for infants in the 1830s.

The systematisation is the contribution

What Froebel added is not any individual fold. Nearly all of them are older, and several are common to the Japanese tradition, which raises a transmission question nobody has settled.

What he added is the sequence: an ordered body of folds, with stated purposes, taught the same way to many children in many places, written down. That is the first time paper folding is treated as a body of technique with an order to it rather than as a set of things some people can do.

And a body of technique with an order to it is the ancestor of the diagram sequence, which is the ancestor of the crease pattern — which is the object this whole site is built around.

Folding a strip into thirdsA guess, and then halving. Each fold moves the mark to the midpoint of one of the two pieces, and each fold halves the distance to the exact division — so the error falls geometrically from whatever the first guess was. The sequence of halvings is read off the binary expansion of the fraction rather than chosen, and the halving of the error is asserted rather than observed.guessoff by 0.16667fold 1off by 0.08333fold 2off by 0.04167fold 3off by 0.02083fold 4off by 0.01042fold 5off by 0.00521fold 6off by 0.00260solid mark — the fold is aiming at 1/3; dashed — at where 1/3 has gonehalving word R L — period 2, from 2^2 − 1 = 3 × 1every fold halves the error exactly, so 6 folds divide it by 64
Fig. 3 Where the systematic approach ends up. Fujimoto’s method reaches an exact odd fraction of a strip by a repeating word of folds, converging on the exact point rather than approximating it — the error halves at every fold, and the generator asserts it to a part in a million million.

What the kindergarten spread

The distribution matters as much as the content, and it is why this lineage is visible in the record when others are not.

Kindergartens spread across Europe, into Britain and America, and — the part that closes a loop — into Japan, where they arrived in the 1870s as part of a wholesale importation of Western educational method. So a European systematisation of paper folding was taught in Japanese schools within a few decades of its invention, in a country with its own folding tradition.

The two traditions are therefore not independent after about 1880. Anybody trying to date a fold by which tradition it belongs to is working in a period when they had already met.

Where the material came from is not settled

One honest gap, and it is the kind the record essay is about.

Several of Froebel’s folds are identical to folds in the Japanese tradition — the same boat, the same box, the same sequence of preliminary folds — and the two traditions were not in contact when the kindergarten material was assembled. Three explanations are available: independent invention, transmission by a route nobody has documented, or the folds being in some sense forced by the square.

The third is not as weak as it sounds. A square has a small number of natural first moves, and a folder exploring them systematically covers the same ground as any other folder doing the same. The same argument accounts for one vertex being found repeatedly in the modern subject.

Nothing in the record settles which. It is listed here because an essay about a lineage should say where the lineage’s own sources are unknown.

Halving is where a syllabus naturally stops

There is a reason the folds of knowledge look the way they do, and it is arithmetic rather than pedagogy.

Repeated halving is the one operation a folder gets for free: bring an edge to an edge, and the sheet is in two. Do it again and there are four, then eight, then sixteen. So a syllabus built from what is easy produces halves, quarters and eighths — every denominator a power of two — and stops.

Thirds are the wall. There is no sequence of edge-to-edge folds that produces an exact third, because no power of two is divisible by three, and a child who has divided a square into sixteenths has no route at all to a ninth.

Folding a strip into fifthsA guess, and then halving. Each fold moves the mark to the midpoint of one of the two pieces, and each fold halves the distance to the exact division — so the error falls geometrically from whatever the first guess was. The sequence of halvings is read off the binary expansion of the fraction rather than chosen, and the halving of the error is asserted rather than observed.guessoff by 0.10000fold 1off by 0.05000fold 2off by 0.02500fold 3off by 0.01250fold 4off by 0.00625fold 5off by 0.00313fold 6off by 0.00156fold 7off by 0.00078fold 8off by 0.00039solid mark — the fold is aiming at 1/5; dashed — at where 1/5 has gonehalving word R R L L — period 4, from 2^4 − 1 = 5 × 3every fold halves the error exactly, so 8 folds divide it by 256
Fig. 4 What lies immediately past the syllabus. Reaching an exact third needs a different move — a guess, corrected by a repeating word of folds whose error halves each time — and it converges on the exact point rather than approximating it. Nothing in the kindergarten material goes here.

That boundary is exactly where the interesting mathematics of folding starts, and the material stops just short of it. The method that crosses it is a twentieth-century one.

The wall is one axiom, not one operation

The thirds boundary is worth naming precisely, because the syllabus was one crease away from crossing it and the crease is not an exotic one.

Every fold in the folds of knowledge brings an edge onto an edge — corner to corner, side to side, a crease onto a crease. That is one operation from the modern list of seven, the angle bisector, and by itself it generates exactly the dyadic rationals: halves, quarters, eighths, and nothing whose denominator carries a factor of three.

The next operation on the list is the fold through two marked points, and it changes the answer immediately. Take a square, halve one edge to get its midpoint — one edge-to-edge fold, entirely within the syllabus. Now crease from the opposite corner through that midpoint, and crease the diagonal. The two lines are y=x/2y = x/2 and y=1−xy = 1-x, and they cross at

x=23,y=13x = \tfrac23, \qquad y = \tfrac13

An exact third, from a midpoint the syllabus already has, in two creases.

Which the material never takes

So the wall was never arithmetic. Powers of two are not divisible by three, and that is a true statement about halving and not a statement about folding — the ceiling is the operation the syllabus restricted itself to, not the medium.

Froebel’s folds are all bisections and superpositions because those are the ones a five-year-old can perform without a marked point to aim at. Bringing an edge onto an edge is self-aligning: the paper tells the child when it is right. Creasing through two points is not — the fold has to be swung until it passes through both, which is a judgement rather than an alignment, and it is exactly the operation the modern axiom list opens with.

So the kindergarten’s ceiling is a pedagogical choice showing up as an algebraic one. The syllabus is confined to self-aligning folds, self-aligning folds give the dyadics, and a third is one non-self-aligning crease away.

That reframes the boundary and makes it more interesting than a wall. The material stops where a child stops being able to check their own work, which is the right place for a curriculum to stop and is not where the mathematics stops. And it explains why the iterative methods that do reach a third are twentieth-century: they get exactness back by making a bad guess and correcting it, which restores the self-checking property at the cost of never finishing in finitely many folds.

The evidence is unusually good

It is worth remarking, in a field where most claims rest on a single source, that this one does not.

Froebel’s material is printed, dated, institutional, and survives in quantity, because it was a curriculum distributed to schools. There are training manuals, there are teachers’ guides, there are children’s exercise books. The 1838 date is not an inference from a poem.

Folding a strip into fifthsA guess, and then halving. Each fold moves the mark to the midpoint of one of the two pieces, and each fold halves the distance to the exact division — so the error falls geometrically from whatever the first guess was. The sequence of halvings is read off the binary expansion of the fraction rather than chosen, and the halving of the error is asserted rather than observed.guessoff by 0.10000fold 1off by 0.05000fold 2off by 0.02500fold 3off by 0.01250fold 4off by 0.00625fold 5off by 0.00313fold 6off by 0.00156fold 7off by 0.00078fold 8off by 0.00039solid mark — the fold is aiming at 1/5; dashed — at where 1/5 has gonehalving word R R L L — period 4, from 2^4 − 1 = 5 × 3every fold halves the error exactly, so 8 folds divide it by 256
Fig. 5 The evidence is unusually good, and so is the method: fifths from a poor guess in eight folds. Each fold divides the remaining error, which is why a child’s first crease being crooked does not matter to where the last one lands.

That is a real asymmetry in the record and it distorts the picture. The European educational lineage looks better documented than the Japanese recreational one because schools generate paper and households do not, and reading that as a difference in antiquity or importance would be exactly the error the record essay warns about.

The box, and what a folded object knows

The folds of life deserve one paragraph of defence, because a modern reader files them as the babyish category and they are not.

A folded box is a three-dimensional object produced from a flat sheet by a stated sequence, and every property it has — that the sides meet, that the corners close, that it stands — is a consequence of the plane geometry of the pattern. Nothing is glued, nothing is measured, and nothing is adjusted at the end. A child who folds a box has, without being told, produced an instance of the central fact of this subject: that a flat sheet’s crease pattern determines a solid.

Folding a strip into eleventhsA guess, and then halving. Each fold moves the mark to the midpoint of one of the two pieces, and each fold halves the distance to the exact division — so the error falls geometrically from whatever the first guess was. The sequence of halvings is read off the binary expansion of the fraction rather than chosen, and the halving of the error is asserted rather than observed.guessoff by 0.40909fold 1off by 0.20455fold 2off by 0.10227fold 3off by 0.05114fold 4off by 0.02557fold 5off by 0.01278fold 6off by 0.00639fold 7off by 0.00320fold 8off by 0.00160fold 9off by 0.00080fold 10off by 0.00040solid mark — the fold is aiming at 1/11; dashed — at where 1/11 has gonehalving word R L R R R L R L L L — period 10, from 2^10 − 1 = 11 × 93every fold halves the error exactly, so 10 folds divide it by 1024
Fig. 6 The box, and what a folded object knows: a division into elevenths reached in seven folds. Nothing about the syllabus needed a number like this, and nothing about the method cares which number is asked for.

What the folds of knowledge did not reach

The syllabus is a geometry syllabus and its ceiling is worth stating, because it is the same ceiling the whole classical tradition has.

Everything in the folds of knowledge is a bisection, a superposition, or a division into equal parts by repeated halving. All of that is within the reach of a straightedge and compass; folding is being used as a convenient instrument rather than a more powerful one.

The thing that makes folding genuinely stronger than the compass — that a single fold can solve a cubic, and therefore trisect and double the cube — is nowhere in the material, and was not known to anybody at the time. Beloch’s paper is a century later.

So the kindergarten used folding for exactness and never discovered its reach. That is not a criticism; it is a nice illustration that a tool’s practical advantages and its theoretical ones are found separately and in either order.

What the syllabus did to the subject

The long consequence is easy to miss because it is institutional rather than intellectual.

Teaching folding to every child in a school system produces, within a generation, a large population who have folded systematically and can be assumed to know what a crease is. That is the precondition for a literature — for a book that says “fold a preliminary base” and expects to be understood, and eventually for a diagram notation that assumes a shared repertoire of moves.

Before that, every account has to start from nothing. After it, an author can build. The kindergarten did not make the subject mathematical, but it made the audience that a mathematical treatment could later be addressed to, which is a less glamorous contribution and possibly a larger one.

The idealisation, named

The figures here draw exact constructions and real children fold real paper, which has thickness, stretch, and a crease that is not a line.

A bisection made by bringing corner to corner is exact in the model and is accurate to whatever the folder can manage in the paper. For the kindergarten’s purposes that is entirely adequate — the point being taught is that the construction is the bisection, not that the child’s crease is perfect — but a figure here showing a crease as a mathematical line is showing the model rather than the classroom.

A curriculum is a claim about what is teachable

The last thing worth taking from the material is methodological, and it applies to this site as much as to a kindergarten.

A syllabus is an assertion that a body of knowledge has an order — that this can be understood before that, and that starting elsewhere fails. Froebel’s order is halving before pattern before object, exactness before decoration, and the sheet before the solid. Whether or not that is the right order, committing to one is what turns a set of tricks into something transmissible.

This site makes the same kind of assertion in a different notation. Every essay carries a rung on a ladder and a list of what it presumes, and the ladder pages exist because a deep argument is unreadable without a path into it. The claim being made is identical in form to Froebel’s: not that these are the facts, but that this is the order in which they become intelligible.

The difference is that the kindergarten’s order was tested on several million children and this site’s has not been tested on anybody.

Where this goes next

A schoolteacher’s theorem is the next rung, and it is the same idea a century and a half on: Kazuo Haga, teaching in a Japanese classroom, folding a corner to a midpoint and finding exact thirds — an amateur result in the best sense, which turned into a small research programme.

The surprising connection is worth ending on. The two things this subject most needed in order to become cumulative are a notation and a systematic body of technique, and neither came from the art. The notation came from a single folder’s decision about how to draw a dashed line; the systematic body came from a man designing a syllabus for infants. Neither was trying to build a mathematical subject.

Where this goes next

the midpointwhat comes outcrease on the left edge 3/8crease on the right edge 7/8the folded edge crosses at2/3exact, and a trisectionthe corner is placed by folding, not by measuring — which is why theresult is exact A schoolteacher's theorem Kazuo Haga folded a corner of a square to the midpoint of the far side and found exact thirds. The construction needs one fold, no measurement and no compass, the numbers that come out are exactly 3/8, 7/8 and 2/3, and it was found by a biology teacher looking for something to do with a classroom. 3 equal partsestimated by eyethe left-hand divisions are exact — a consequence of the fold, not of carethe right-hand ones are a guess, and the error compoundsvalleymountain Dividing without measuring A square can be divided into any whole number of equal parts by folding alone — exactly, with no ruler, and with no error to accumulate. The construction is one fold and a theorem nobody expected. Paper is made in ChinaPaper reaches JapanPaper is made in EuropeFolded paper is used ceremonially in Japan400 yrPaper is folded for amusement in Japan980 yrThe thousand cranes897 yrThe pajarita is folded in Spain293 yrPaper folding is taught as geometryOne fold solves a cubicThe diamond pattern in a crushed cylinderThe conditions at a flat-foldable vertexThe dashed-and-dotted diagram notationThe Miura foldA five-pointed star from one straight cutAny straight-line drawing, from one straight cutyear of the source500100015002000the date generally giventhe oldest source that says somedian overrun 201.5 years Nothing here is as old as it sounds Paper folding is described everywhere as an ancient art. The oldest surviving book of it was printed in 1797, the oldest reference to folding for amusement is from 1680, and the median claim in this subject is dated two centuries before anything that attests it.

What this makes readable

Essays that name this one as a prerequisite.

Named alongside this one

Essays reaching for the same objects. Nobody chose these; they are what the concept index makes visible.

What links here

The 8 essays that link to this one and share the most of its objects, of 10 that link here.

The objects this essay names

Each one links to every other essay that touches it.

Constructive proofExact divisionFroebelFujimoto's methodPedagogy