The 8 motions of a regular 4-gon
group is one function. Everything below came out of it during this
build, at parameters taken from the essays rather than invented for this page — so a figure
here is the same figure a reader meets in an essay, and if the generator changes, this page
changes with it.
With nothing chosen
16 colourings in 6 classes
Colourings left alone by each motion, and their average
Every relabelling of a 4-gon's corners, and the 8 that are motions
The 8 motions of a regular 4-gon
The composition table of the 8 motions
What it checks while it draws
Collected by running the family and recording what it asserted, not written here. The count is how many separate times the claim was put to the test while these drawings were made.
- the hex ball of radius 0 ×13
- the king ball of radius 0 ×13
- the square ball of radius 0 ×13
- the pieces of Z/2 ∗ Z/2 outside radius 0 that reach the outer shell do not change when it moves out ×10
- a subgroup of 1 divides the group of 8 ×8
- the ball of radius 0 in F₂ has the size the formula gives ×8
- the ball of radius 0 in Z has the size the formula gives ×8
- the ball of radius 0 in Z² has the size the formula gives ×8
- D3 has an element of order 2 ×7
- D3: every solution but e has order exactly 2 ×7
- D3: the solutions of g^2 = e are a multiple of 2 ×7
- the ball of radius 0 in Z/2 ∗ Z/2 has the size the formula gives ×7
- the ball of radius 0 in Z³ has the size the formula gives ×7
- D4: a class's size times its centraliser is the group ×6
- D4: a group that does not commute reaches at most 5/8 ×6
- D4: the centralisers of one class add up to the group ×6
- D4: the centre is at most a quarter of the group ×6
- D4: the commuting pairs are the size times the number of classes ×6
- the solutions of x^1 = e are a multiple of 1 ×6
- a subgroup of order 8 exists ×5
- A4 has an element of order 2 ×5
- A4: every solution but e has order exactly 2 ×5
- A4: the solutions of g^2 = e are a multiple of 2 ×5
- the pieces of F₂ outside radius 0 that reach the outer shell do not change when it moves out ×5
- the pieces of Z × Z/2 outside radius 2 that reach the outer shell do not change when it moves out ×5
- the pieces of Z outside radius 2 that reach the outer shell do not change when it moves out ×5
- the pieces of Z/6 outside radius 0 that reach the outer shell do not change when it moves out ×5
- the pieces of Z² outside radius 2 that reach the outer shell do not change when it moves out ×5
- the pieces of Z³ outside radius 0 that reach the outer shell do not change when it moves out ×5
- every motion sending corner 1 to 0 is one of 2 ×4
- P2: a class's size times its centraliser is the group ×4
- P2: a group that does not commute reaches at most 5/8 ×4
- P2: the centralisers of one class add up to the group ×4
- P2: the centre is at most a quarter of the group ×4
- P2: the commuting pairs are the size times the number of classes ×4
- the Pauli group on 1 has 4^1 + 1 classes ×4
- the Pauli group on 1 has a centre of two ×4
- the Pauli group on 2 has 2·4^2 elements ×4
- S3: the classes are the partitions of 3 ×3
- S4 has the number of elements it should ×3
- S4: a class's size times its centraliser is the group ×3
- S4: a group that does not commute reaches at most 5/8 ×3
- S4: the centralisers of one class add up to the group ×3
- S4: the centre is at most a quarter of the group ×3
- S4: the commuting pairs are the size times the number of classes ×3
- the elements reachable in 1 steps are exactly those at distance 1 or less ×3
- the radius runs to between 3 and 8 ×3
- A4 has the number of elements it should ×2
- A4: a class's size times its centraliser is the group ×2
- A4: a group that does not commute reaches at most 5/8 ×2
- A4: the centralisers of one class add up to the group ×2
- A4: the centre is at most a quarter of the group ×2
- A4: the commuting pairs are the size times the number of classes ×2
- colouring 1100: orbit times stabiliser is the group ×2
- e leaves 16 colourings alone, one per cycle coloured freely ×2
- exactly 8 of the 24 relabellings preserve every distance ×2
- m₁ leaves 8 colourings alone, one per cycle coloured freely ×2
- m₂ leaves 4 colourings alone, one per cycle coloured freely ×2
- m₃ leaves 8 colourings alone, one per cycle coloured freely ×2
- m₄ leaves 4 colourings alone, one per cycle coloured freely ×2
- r leaves 2 colourings alone, one per cycle coloured freely ×2
- r² leaves 4 colourings alone, one per cycle coloured freely ×2
- r³ leaves 2 colourings alone, one per cycle coloured freely ×2
- S4 has an element of order 2 ×2
- S4: every solution but e has order exactly 2 ×2
- S4: the solutions of g^2 = e are a multiple of 2 ×2
- the classes hold all 16 colourings between them and share none ×2
- the drawing of Z/2 ∗ Z/2 shows as many outward pieces as are counted ×2
- the radius is between 2 and 6 ×2
- a class's size times the relabellings fixing one graph is n! ×1
- a class's size times what commutes with one member is the group ×1
- a colouring gives one colour per corner ×1
- a finite group has nothing left outside a large enough ball ×1
- a pair names two different corners ×1
- a point reached in d steps lies inside d·P ×1
- a subgroup whose blocks differ has an element that shows it ×1
- a tree walk is back after two steps with chance 1/4 ×1
- a tuple no turn changes repeats one element ×1
- a turned tuple still has product e ×1
- above the threshold the diagonal time is the minimum, below it more ×1
- all four pieces are non-empty ×1
- all graphs are drawn for at most four points ×1
- and a finite stretch between the cuts ×1
- and divides the index ×1
- and does not fall away ×1
- and it is already below three fifths ×1
- and nearer at the larger radius ×1
- and no element is in two of them ×1
- and none of them is the identity, which would generate nothing ×1
- and that does not change as the ball grows ×1
- and the second subgroup's are not — which is what makes it the control ×1
- and the two maps are genuinely different graphs ×1
- and the whole group is reached ×1
- and there is a multiple of p of them ×1
- at most 20,000 tuples are enumerated ×1
- axis + one diagonal: |B(r)|/r² is near the polygon's area ×1
- axis steps: |B(r)|/r² is near the polygon's area ×1
- between one and four groups this family knows ×1
- between one and six things are tabled ×1
- between one and three generating sets among square, hex, king, knight ×1
- between one and three generators are named ×1
- between one and three groups this family knows ×1
- between three and eight groups this family knows ×1
- between two and four generating sets ×1
- both maps are of the same group ×1
- column e does too ×1
- column m₁ does too ×1
- column m₂ does too ×1
- column m₃ does too ×1
- column m₄ does too ×1
- column r does too ×1
- column r² does too ×1
- column r³ does too ×1
- corner 1: orbit times stabiliser is the group ×1
- corners times the rotations holding one still is every rotation ×1
- diagonal 1–3: orbit times stabiliser is the group ×1
- e preserves every distance between corners ×1
- each step's size divides the last ×1
- edge 1–2: orbit times stabiliser is the group ×1
- edges times the rotations holding one still is every rotation ×1
- every block is the size of the subgroup ×1
- every class has a size dividing the group's ×1
- every composition of two motions is again one of the motions ×1
- every corner of the cube can be carried to every other ×1
- every corner of the dodecahedron can be carried to every other ×1
- every corner of the icosahedron can be carried to every other ×1
- every corner of the octahedron can be carried to every other ×1
- every corner of the tetrahedron can be carried to every other ×1
- every edge of the cube can be carried to every other ×1
- every edge of the dodecahedron can be carried to every other ×1
- every edge of the icosahedron can be carried to every other ×1
- every edge of the octahedron can be carried to every other ×1
- every edge of the tetrahedron can be carried to every other ×1
- every element found is accounted for in the shells ×1
- every element of the tree of triangles is placed ×1
- every element sits in exactly one shell ×1
- every face of the cube can be carried to every other ×1
- every face of the dodecahedron can be carried to every other ×1
- every face of the icosahedron can be carried to every other ×1
- every face of the octahedron can be carried to every other ×1
- every face of the tetrahedron can be carried to every other ×1
- every generator names an element of the group ×1
- every group drawn has a picture in the plane ×1
- every named element is in the group ×1
- every point of the box is reached ×1
- every reduced word up to the stated length is listed ×1
- every ring's size divides the length ×1
- every set has a non-empty edge inside it ×1
- every shorter word is in exactly one piece of the a split ×1
- every shorter word is in exactly one piece of the b split ×1
- every tuple built has product e ×1
- F₂ multiplies its ball by more than two at each step ×1
- F₂'s count rises at every radius ×1
- faces times the rotations holding one still is every rotation ×1
- graphs on three to five points ×1
- in a line a second cut leaves two outward pieces ×1
- in a tree the second cut adds outward pieces ×1
- king's moves: |B(r)|/r² is near the polygon's area ×1
- knight's moves: |B(r)|/r² is near the polygon's area ×1
- m₁ preserves every distance between corners ×1
- m₂ preserves every distance between corners ×1
- m₃ preserves every distance between corners ×1
- m₄ preserves every distance between corners ×1
- m₅ preserves every distance between corners ×1
- m₆ preserves every distance between corners ×1
- most of a ball in F₂ is its outermost shell ×1
- multiplying on the left carries every edge to an edge ×1
- no missed point is more than three polygon-steps inside the edge ×1
- no point is reached in fewer steps than its norm ×1
- no two of the motions do the same thing to the corners ×1
- no word is in both pieces of the a split ×1
- no word is in both pieces of the b split ×1
- on the tree the distance grows at about half a step per step ×1
- one block for each place in the orbit ×1
- one rotation for each way of putting down a directed edge ×1
- Q8: a class's size times its centraliser is the group ×1
- Q8: a group that does not commute reaches at most 5/8 ×1
- Q8: the centralisers of one class add up to the group ×1
- Q8: the centre is at most a quarter of the group ×1
- Q8: the commuting pairs are the size times the number of classes ×1
- r preserves every distance between corners ×1
- r² preserves every distance between corners ×1
- r³ preserves every distance between corners ×1
- r⁴ preserves every distance between corners ×1
- r⁵ preserves every distance between corners ×1
- row e contains every element exactly once ×1
- row m₁ contains every element exactly once ×1
- row m₂ contains every element exactly once ×1
- row m₃ contains every element exactly once ×1
- row m₄ contains every element exactly once ×1
- row r contains every element exactly once ×1
- row r² contains every element exactly once ×1
- row r³ contains every element exactly once ×1
- so the subgroup's size times the number of blocks is the group's size ×1
- some group in the table does not reach the identity ×1
- some pair of motions gives a different result in each order ×1
- the average number left alone is the number of classes ×1
- the ball is drawn for the groups with a two-dimensional picture ×1
- the ball keeps growing ×1
- the bars are drawn for at most 24 elements ×1
- the block a product lands in does not depend on which elements were picked ×1
- the blocks between them cover the group exactly once ×1
- the bound is drawn for a group that does not commute ×1
- the box runs to between 6 and 16 ×1
- the chance falls and stays above one half ×1
- the classes are drawn for at most 64 colourings ×1
- the classes hold every element once ×1
- the classes hold every graph exactly once ×1
- the commutators generate a subgroup of the group they came from ×1
- the conjugacy classes account for every element exactly once ×1
- the corners are coloured in between 2 and 4 colours ×1
- the derived subgroup is carried into itself by every conjugation ×1
- the drawing has extent in both directions ×1
- the drawing of F₂ shows as many outward pieces as are counted ×1
- the drawing of Z × Z/2 shows as many outward pieces as are counted ×1
- the drawing of Z shows as many outward pieces as are counted ×1
- the drawing of Z² shows as many outward pieces as are counted ×1
- the drawing reaches at least two steps past the cut ×1
- the dual's corners point at the faces in one of its orientations ×1
- the edge of a ball in F₂ stays a large share of it ×1
- the edge of a ball in Z² is a falling share of it ×1
- the elements are closed under composition ×1
- the estimate falls as the distance grows ×1
- the even symmetries of five letters have no normal subgroup in between ×1
- the exhaustive search is drawn for at most 5 corners ×1
- the first has as many blocks as its index ×1
- the first subgroup is named by element labels ×1
- the first subgroup's left and right blocks are the same blocks ×1
- the four pieces and the identity account for every word ×1
- the generating set is one this family knows ×1
- the generators are distinct ×1
- the grid is drawn for at most twelve elements ×1
- the group is a dihedral group D3 to D8 or a named permutation group ×1
- the group is one of Z, Z2, Z3, F2, D, T, ZZ2, C6 ×1
- the group is one this family knows ×1
- the group is the dihedral one or the cyclic one ×1
- the groups are among Z, Z2, Z3, F2, D, T, ZZ2, C6 ×1
- the groups drawn do not all grow alike ×1
- the identity is a class of its own ×1
- the knight misses points inside its polygon ×1
- the labelled trees number n to the power n − 2, as Cayley's formula says ×1
- the lattice counts agree with direct enumeration ×1
- the lattices' roots approach 1 ×1
- the m₁ and m₂ generate the whole group ×1
- the measured rate matches 2√(k − 1)/k ×1
- the motions sending the thing to one place are one coset of the stabiliser ×1
- the named elements are closed under composition ×1
- the outer sphere is (k − 2)/(k − 1) of a large ball ×1
- the outermost shell of Z is a falling share of the ball ×1
- the outermost shell of Z² is a falling share of the ball ×1
- the permutations are of between two and five letters ×1
- the polygon has between 3 and 8 corners ×1
- the r and m₁ generate the whole group ×1
- the r generate the whole group ×1
- the radius removed is between 0 and 3 ×1
- the removed radius runs to between 2 and 5 ×1
- the rings hold every tuple once ×1
- the second is too ×1
- the series never grows ×1
- the sets run to size parameter 4 to 12 ×1
- the share of the polygon reached does not fall as r grows ×1
- the singletons are exactly the elements whose p-th power is e ×1
- the solid is one of the five ×1
- the spread grows like a power near one third ×1
- the subgroup is named by between one and eight element labels ×1
- the Sylow subgroups are exactly the conjugates of one of them ×1
- the three-dimensional count agrees with enumeration ×1
- the translate's centre is inside the drawing ×1
- the tree's root approaches √3/2 ×1
- the trivial subgroup and the whole group are both among them ×1
- the tuples are drawn when there are at most 64 of them ×1
- the tuples with product e number |G| to the power n − 1 ×1
- the two subgroups nobody has to look for are among the ones found ×1
- the view is one the family draws ×1
- the word e lands on e ×1
- the word length is never more than three above the norm ×1
- the word m₁ lands on m₁ ×1
- the word m₁·m₂ lands on r³ ×1
- the word m₁·m₂·m₁ lands on m₄ ×1
- the word m₁·m₂·m₁·m₂ lands on r² ×1
- the word m₁·r lands on m₄ ×1
- the word m₁·r lands on m₅ ×1
- the word m₁·r lands on m₆ ×1
- the word m₁·r·m₁ lands on r⁴ ×1
- the word m₁·r·m₁ lands on r⁵ ×1
- the word m₁·r·r lands on m₄ ×1
- the word m₁·r·r lands on m₅ ×1
- the word m₂ lands on m₂ ×1
- the word m₂·m₁ lands on r ×1
- the word m₂·m₁·m₂ lands on m₃ ×1
- the word r lands on r ×1
- the word r·m₁ lands on m₂ ×1
- the word r·r lands on r² ×1
- the word r·r·m₁ lands on m₃ ×1
- the word r·r·r lands on r³ ×1
- the word r·r·r·m₁ lands on m₄ ×1
- the word r·r·r·r lands on r⁴ ×1
- the word r·r·r·r·r lands on r⁵ ×1
- the words run to between 3 and 8 letters ×1
- the words strictly inside the bound are counted ×1
- their number leaves remainder 1 on division by p ×1
- they differ in something a reader can see — the number of steps across, or the number of arrows ×1
- three or four increasing radii up to 24 ×1
- two generating sets are compared ×1
- two motions with the same target differ by a motion holding the thing still ×1
- two of Z, F2, T and D are compared ×1
- under a prime turn every ring has 1 or p tuples ×1
- Z × Z/2 settles on 0, 1 or 2 ×1
- Z does not ×1
- Z settles on 0, 1 or 2 ×1
- Z/2 ∗ Z/2 does not ×1
- Z/2 ∗ Z/2 settles on 0, 1 or 2 ×1
- Z/2 ∗ Z/3's count rises at every radius ×1
- Z/6 settles on 0, 1 or 2 ×1
- Z² does not ×1
- Z² settles on 0, 1 or 2 ×1
- Z³ does not ×1
- Z³ settles on 0, 1 or 2 ×1
Where it is called
Every figure on this list is drawn by the same rule, so a change to the rule changes all of them at once. That is why the list is published.
Colourings nobody can tell apart
Sixteen ways to colour four corners in two colours, and only six of them are genuinely different. The count can be got by pooling the sixteen — or by never forming a single class and instead averaging how many colourings each motion leaves untouched.
AlgebraEight ways to leave a square alone
A square can be picked up and put back so that nothing looks different. There are exactly eight ways to do it, and the number is not asserted here — it is what a search through all twenty-four relabellings of the corners comes back with.
AlgebraFive-eighths of the pairs, and no more
Pick two symmetries of a square at random and do them in both orders: forty times in sixty-four the result is the same. No group that fails to commute does better. The reason is a count of pairs that turns into a count of conjugacy classes, and a two-line argument about the centre that caps the answer at five-eighths — reached by the square and the quaternions, approached from above by nothing, and approached from below by groups that commute a little more than half the time.
AlgebraHow fast the ball fills
Count the elements within r steps of doing nothing. The count grows like a polynomial in some groups and like a power of three in others, the distinction survives every change of generating set, and which polynomial degrees are possible is a theorem nobody expected.
AlgebraHow rarely a walk on a group comes home
Walk at random on the picture of a group, one generator at a time, and ask for the chance of standing at the start after 2n steps. On the line, the plane and three-dimensional space it falls like a power of n. On the tree that pictures the free group it falls by the factor √3/2 every step, exponentially. Kesten proved in 1959 that this is no accident of two examples: the chance falls exponentially exactly when the group's balls are mostly boundary, so a probabilistic rate and a geometric ratio are the same measurement.
AlgebraNecklaces made of symmetries
Lagrange's theorem says a subgroup's size divides the group's, and the converse is false. One piece of the converse is true: every prime that divides the size is the order of some element. The proof threads the group's own elements onto a necklace whose product is nothing, turns it, and counts — the argument that proved Fermat's little theorem with beads, with the beads replaced by motions.
AlgebraThe blocks a subgroup cuts out
Take any part of a group that is closed under composition, and it slices the whole group into blocks of its own size that do not overlap. Everything Lagrange's theorem says is arithmetic about that picture — and whether the blocks can be multiplied is a separate question with a surprising answer.
AlgebraThe crossings that will not come out even
Draw a rearrangement as strings from one row of pegs to another and count where they cross. The count depends on how the strings are drawn; whether it is odd or even does not, and that single bit is what makes determinants exist and a sliding puzzle unsolvable.
AlgebraThe edge that is as big as the ball
In a lattice the boundary of a large ball is a negligible fraction of it. In a tree it is two thirds of it at every size — and that single ratio, not the group's size, is what decides whether a set can be cut into pieces and reassembled into two copies of itself.
AlgebraThe group drawn as a map
A multiplication table says everything about a group and shows nothing; lay the same information out as one dot per element and one arrow per generator, and multiplying becomes walking, distance becomes a word length, and the group acquires a shape.
AlgebraThe group that will not come apart
Solving an equation by radicals means building a tower of roots, and a tower of roots corresponds to a chain of subgroups with abelian steps. For the general equation of degree five that chain would have to descend through a group of sixty elements with no normal subgroup in it — so there is no formula, and the obstruction is a finite object that can be written out.
AlgebraThe lattice that runs the other way
The symmetries of a polynomial's roots form a group, and the fields between the bottom and the top form a lattice. The two are the same picture, one of them turned over — a bigger group of symmetries fixes less, so it names a smaller field.
AlgebraThe polygon a lattice becomes from far away
Walk the grid of whole-number points with a fixed set of moves and the places reachable in r moves fill a shape. With axis steps it is a diamond, add a diagonal and it is a hexagon, move like a knight and it is a ragged thing full of holes — which, seen from far enough away, is an octagon exactly. The generators decide the polygon, and the polygon decides the count.
AlgebraThe shape a random ball grows into
Give every road of the square grid a random travel time and ask what can be reached from one point in time t. The region is ragged, and rescaled it converges to a fixed convex shape — but which shape is unknown for every natural law. Computing it shows a curve within a few per cent of a circle for continuous travel times, a flat side where fast roads percolate along a diagonal, a time per step that is still drifting at a hundred and twenty-eight steps, and fluctuations that grow like the distance to the power one third rather than one half.
AlgebraTwenty-four ways to set a cube down
Count the rotations of a cube from its corners and the answer is eight times three. Count from its edges and it is twelve times two; from its faces, six times four. Three different pictures give one number because each count is the same theorem — the places a thing can go, times the motions that leave it where it is — and the same theorem splits Cayley's sixteen trees into twelve and four and proves that a group of eight has a centre.
AlgebraWhat is left when the middle is taken out
Cut a finite piece out of a group's picture and count the parts of what remains that run off forever. The integers leave two, the plane one, a tree more with every cut — and no group anywhere leaves exactly three, because a third end is always the first of infinitely many.