The quaternion multiplication table, from i² = j² = k² = ijk = −1
quaternion 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.
At its defaults
show: "rotate"
show: "double"
show: "algebras"
show: "so4"
show: "hopf"
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 units 1 and 2 lie on exactly one line ×21
- 24 of the 64 ordered pairs of units fail to commute ×1
- 3 and 5 are each a sum of three squares and 15 is not, so no three-square identity exists ×1
- a labelling exists on which every line reads in its own multiplication order ×1
- a left multiplication and a right multiplication commute ×1
- a product of whole octonions is whole ×1
- a product of whole quaternions is whole ×1
- a quaternion and its negative perform the same rotation ×1
- all seven units are placed ×1
- and every non-zero octonion still has an inverse ×1
- and every one of them has length one ×1
- and every one of them has squared length two ×1
- and every point of a fibre lies over the same point of the sphere ×1
- and every root sees the same profile of neighbours as every other ×1
- and in the ring with halves it is closer than one, which is a division algorithm ×1
- and its determinant is one, so it turns rather than reflects ×1
- and most of every circle is still drawn, so the breaks read as crossings ×1
- and no two circles touch ×1
- and none of them has two arguments the same, which is alternativity ×1
- and one hundred and twenty-eight have every coordinate a half ×1
- and one is its negative ×1
- and the five classes account for every root ×1
- and the other six are straight ×1
- and the quotient that achieves it is one of the halves ×1
- and the rotations they perform number twelve, two units to each ×1
- and the trace of its square gives the same two angles again ×1
- and the twenty-four sit inside the hundred and twenty five times over ×1
- and the two-square identity is the same statement in the plane ×1
- and their real parts take nine values ×1
- and they are twenty-four different quaternions ×1
- and they are two hundred and forty different vectors ×1
- and tilts within a right angle ×1
- between two and eight circles are drawn ×1
- conjugating a pure quaternion leaves it pure ×1
- doubling the complex numbers gives Hamilton's own multiplication ×1
- each circle is sampled at between 60 and 400 points ×1
- each composite is a rotation rather than a reflection ×1
- each unit has eight others at distance one, which is the 24-cell's vertex figure ×1
- every icosian is counted once ×1
- every number up to the bound is a sum of four squares ×1
- every pair of circles is linked exactly once ×1
- every pair of roots meets at an inner product of −2, −1, 0, 1 or 2 ×1
- every point of a fibre is on the unit three-sphere ×1
- exactly one of the seven lines is the drawn circle ×1
- fifty-six roots stand at sixty degrees to any given one ×1
- i j k = −1, which is the rule the rest follows from ×1
- in the whole-coordinate ring the quotient is a full unit away, so no remainder is smaller than the divisor ×1
- its centre is exactly plus and minus one ×1
- negating both quaternions gives the same rotation ×1
- negating only one of them does not ×1
- no drawn circle comes near the point the projection removes ×1
- no two vertices project onto one another ×1
- one full turn of the object leaves the quaternion at −1 ×1
- one hundred and twenty-six stand at a right angle ×1
- one of them is the identity ×1
- reflecting one root in another lands on a root ×1
- sixty-four straddle the two halves ×1
- some strand passes behind another, which is what makes the picture a diagram ×1
- some triples of units fail to associate ×1
- the associator changes sign when two of its arguments are swapped ×1
- the base points sit at a latitude strictly inside the poles ×1
- the columns of the map are orthonormal, so lengths are kept ×1
- the eight squares of the product add to the product of the two sums of eight squares ×1
- the first quaternion is four whole numbers no larger than twenty ×1
- the four squares of the product add to the product of the two sums of four squares ×1
- the highlighted line is one of the seven, or none ×1
- the highlighted product is ij, ji, squares, or none ×1
- the icosians number one hundred and twenty ×1
- the left quaternion is four numbers ×1
- the left quaternion is meant to have length one ×1
- the lift really does send i to the base point ×1
- the multiplication is associative on every triple of units ×1
- the numbers that really need four squares are the ones Legendre's condition names ×1
- the octonions are not even associative ×1
- the plane has seven lines ×1
- the quaternion has length one ×1
- the quaternions do not commute ×1
- the reals commute ×1
- the right quaternion is four numbers ×1
- the right quaternion is meant to have length one ×1
- the rotation after one full turn is the identity again ×1
- the second quaternion is four whole numbers no larger than twenty ×1
- the solid has ninety-six edges ×1
- the sums of squares are checked between 20 and 400 far ×1
- the sweep is taken in 8 to 24 steps, a multiple of four ×1
- the system is symmetric under negation ×1
- the tower stops at dimension 1, 2, 4 or 8 ×1
- the trace of the matrix is twice the sum of the cosines of the two angles ×1
- the turn is between a tenth and a half of a full circle ×1
- the two blocks are orthogonal ×1
- the two blocks are shaded or not ×1
- the two orders give different quaternions ×1
- the two rotations move at least one axis a long way apart ×1
- the units are closed under multiplication ×1
- the view is one the family draws ×1
- the view is tilted between minus ninety and ninety degrees ×1
- the view turns by a real angle ×1
- there are twenty-four units ×1
- there are two hundred and forty roots ×1
- they are closed under multiplication ×1
- thirty of them are half-turns ×1
- twenty-four in the last four ×1
- twenty-four of them are the Hurwitz units ×1
- twenty-four roots live in the first four coordinates ×1
- two different axes from i, j, k ×1
- two full turns bring it back to 1 ×1
- two-dimensional inputs give a two-dimensional answer, which is why the plane closes ×1
- which are themselves closed under multiplication ×1
- which is all of them, counted twice by two different rules ×1
Where it is called
Changing this generator changes every figure on this list. That is what makes the list worth publishing rather than keeping in a check script.
A multiplication that remembers the order
Four characters — i² = j² = k² = ijk = −1 — define a multiplication in which ab and ba are different numbers. Everything follows from them, including the fact that a rotation of a solid body has two names, and that two full turns are needed to get one of them home.
AlgebraA rotation of four-space takes two of them
One unit quaternion, conjugating, turns three-space about an axis. Two of them, multiplying from the left and the right, turn four-space — and a rotation of four-space has no axis at all, but two independent angles and two planes it spins in.
AlgebraThe identity that multiplies sums of squares
A sum of two squares times a sum of two squares is a sum of two squares, and the same is true of four and of eight. It is true of three of nothing: 3 and 5 are each a sum of three squares, 15 is not, and one small pair of numbers rules the case out forever.
AlgebraThe integers among the quaternions
The obvious integer quaternions are the ones with whole coordinates, and they are the wrong ones. Sixteen more, with every coordinate a half, have to be let in — and with them comes a division algorithm, twenty-four units instead of eight, and a regular solid that exists only in four dimensions.
AlgebraTwo hundred and forty directions
The quaternions have twenty-four units and they are the vertices of the most symmetric object in four dimensions. Eight dimensions has two hundred and forty of them, and the quaternions turn out to be how they are built — twice over, with a hundred and ninety-two left to explain.
AlgebraWhat is lost at eight
Double the quaternions and division still works, but ab times c and a times bc are no longer the same number. What survives is a weaker law that turns out to be enough for a great deal — and the whole multiplication table fits into a picture of seven lines.