A linear map redrawing the plane
linear-map 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
A determinant counting the 16 spanning trees
A determinant counting 889 pairs of paths that never meet
2 independent rows and 2 independent columns
Rank two, rank one, rank nothing
A whole line arrives at the origin
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.
- and the 5th power agrees with 5 multiplications ×2
- the matrix is square, 2×2 to 7×7 ×2
- a component made of k discs holds exactly k eigenvalues ×1
- a graph with a cycle has at least as many edges as vertices ×1
- a map that keeps both dimensions has a non-zero determinant ×1
- a map that loses a dimension has determinant zero ×1
- a point outside every disc is not an eigenvalue ×1
- a power of the matrix is between 2 and 24 ×1
- a real 2×2 map has two eigenvalues counted with sign ×1
- a skew matrix exponentiates to a rotation, first entry ×1
- a symmetric matrix's eigen-directions are perpendicular ×1
- a touching pair and a free pair were both found ×1
- adding a multiple of one column to another changes nothing ×1
- an ellipse of area past four times the determinant does hold a non-zero point ×1
- an odd number of grid lines between 3 and 9 ×1
- an odd number of grid lines between 3 and 9, so one runs through the point ×1
- an unsymmetric one's are not, which is what the panel is for ×1
- and agrees with the expansion along a row ×1
- and at a root the shifted map crushes something to nothing ×1
- and does not already point along the answer ×1
- and each direction really is left alone ×1
- and every pair running to the other finishes does cross ×1
- and every point at all lands on one line ×1
- and exponentiating the two eigenvalues gives the same matrix ×1
- and in some column disc ×1
- and it is at most the smaller of the two counts ×1
- and it keeps area exactly ×1
- and its discs obey the same count ×1
- and its size is the size of that number ×1
- and its two eigen-directions are genuinely different ×1
- and moves the permanent ×1
- and multiply to the determinant ×1
- and none passes the next one's starting value ×1
- and reaches it on the smallest patch drawn ×1
- and so are their images ×1
- and so does the triple product of the columns ×1
- and the angle shrinks by the ratio of the two eigenvalues ×1
- and the direction it occurs in is an eigen-direction, to within one sample ×1
- and the identity leaves the unit square alone ×1
- and the lowest at the second ×1
- and the lowest is the smaller ×1
- and the shortest is 1/√λ for the larger ×1
- and the shortest is the second ×1
- and the two add to the number of edges ×1
- and the two are a right angle apart ×1
- and there is a direction it sends to the origin ×1
- and where it has got to is what that rate predicts ×1
- any plane containing the top eigenvector reaches the top eigenvalue ×1
- AᵀA is symmetric and has two real eigenvalues ×1
- Av is λv in the first coordinate ×1
- Av is λv in the second ×1
- both eigenvalues are positive, so the level set is a closed curve ×1
- D⁻¹AD has the same eigenvalues as A ×1
- each eigenvalue makes A − λI singular, by a determinant the polynomial never saw ×1
- each eigenvalue of the smaller block sits between two of the larger's ×1
- each Jacobi pair satisfies Av = λv ×1
- each matrix in a row is four finite entries under 8 in size ×1
- each root really is a root ×1
- every choice of struck row gives the same number ×1
- every corner of the image stays inside its panel ×1
- every eigenvalue lies in some row disc ×1
- every pair was classified once ×1
- every plane's maximum lies between the middle and top eigenvalues ×1
- every point drawn satisfies xᵀAx = 1 ×1
- every point of that line lands on the origin ×1
- every start reaches every finish ×1
- every step brings the arrow nearer the dominant direction ×1
- every tree found has one edge fewer than it has vertices ×1
- fourth ×1
- in both coordinates ×1
- no eigenvalue goes down ×1
- one dimension survives and one is lost, and they add to the two started with ×1
- one eigenvalue is strictly the largest ×1
- one entry of v per row ×1
- one variant at a time ×1
- P D P⁻¹ is the map it started from ×1
- reached at the top eigenvector itself ×1
- scaling one column scales the area by the same factor ×1
- second ×1
- so its nullity is the number of independent cycles, counted from the graph itself ×1
- so relative to the patch it falls by two ×1
- so the determinant counts the pairs that never meet ×1
- swapping two columns turns the sign over ×1
- t runs up to at most 40 ×1
- the area factor of e^A is e to the trace of A ×1
- the area factor of the image closes on |det J| as the patch shrinks ×1
- the area multiplier is zero ×1
- the area of the drawn parallelogram is ad − bc ×1
- the array is one the family knows ×1
- the arrow it starts from is a pair of numbers with a length ×1
- the arrow takes at least three steps before it arrives ×1
- the basis is four finite entries no larger than four ×1
- the basis spans an area rather than a line ×1
- the camera angles are within a half turn ×1
- the change of basis is invertible ×1
- the column is scaled by between one and three ×1
- the count holds at every stage of the continuation ×1
- the counting window is between 2 and 8 across ×1
- the crossing pairs are matched exactly by the pairs running to the other finishes ×1
- the curved map is one the family knows ×1
- the determinant is the number of spanning trees, counted ×1
- the determinant of the product is the product of the determinants ×1
- the drawn area is the determinant ×1
- the eigenvalues add up to the trace ×1
- the eigenvalues rise in total by exactly t|v|² ×1
- the ellipse has two positive radii no larger than four ×1
- the expansion along a row agrees with the sum over permutations ×1
- the first matrix of a product is four finite entries under 8 in size ×1
- the flow runs for a positive time of at most 4 ×1
- the flow runs for between 0 and 4 units of time ×1
- the flow starts from between three and twelve points ×1
- the flow's own arrival points are the columns of the series' answer ×1
- the form has two real eigenvalues ×1
- the form is positive in every direction ×1
- the four terms of the expansion add to ad − bc ×1
- the graph is one the family knows ×1
- the highest point sits at the first eigen-direction ×1
- the highest value on the circle is the larger eigenvalue of the symmetric part ×1
- the image line has a length to normalise ×1
- the image of the unit square has the determinant's area ×1
- the incidence matrix's rank is the number of vertices less the number of components ×1
- the iteration runs between 2 and 40 steps ×1
- the lattice is drawn far enough out to fill the counting window ×1
- the lattice is drawn out to between 3 and 12 steps ×1
- the longest image measured on the ellipse is the first stretch ×1
- the longest radius measured on the curve is 1/√λ for the smaller eigenvalue ×1
- the lowest of the plane-maxima is the middle eigenvalue ×1
- the map does not crush the arrow to nothing at any step ×1
- the map does not flatten the circle to a segment ×1
- the map has two real eigenvalues ×1
- the map has two real eigenvalues to build a basis from ×1
- the map is one the family knows ×1
- the maps drawn have different ranks ×1
- the marked point is a complex number as a pair ×1
- the matrix drawn here collapses the plane onto a line ×1
- the matrix handed to the symmetric solver is symmetric ×1
- the matrix is four finite entries no larger than four ×1
- the matrix is four finite entries under 8 in size and is not all zero ×1
- the matrix is square, 2×2 to 6×6, entries under 40 ×1
- the matrix is three rows of three finite entries no larger than four ×1
- the matrix is three rows of three whole numbers under ten ×1
- the measured Jacobian is the analytic one ×1
- the minmax view is for a 3×3 matrix ×1
- the multiple is a non-zero whole number no larger than four ×1
- the number of independent rows and the number of independent columns are one number ×1
- the patch is at most 1.2 across ×1
- the paths counted by hand agree with the binomial ×1
- the permanent of the pattern counts the ways to pick one entry per row and column ×1
- the point is a pair of coordinates ×1
- the points in a window are its area over the determinant, up to the boundary ×1
- the residual falls by four at every halving of the patch ×1
- the roots add to the trace ×1
- the rotation method and the polynomial agree ×1
- the row operation adds one row to a different one ×1
- the row operation leaves the determinant exactly where it was ×1
- the scaling is one positive weight per row ×1
- the second is too ×1
- the series has converged ×1
- the shifted determinant and λ² − (trace)λ + determinant agree ×1
- the signed sum is the determinant ×1
- the smallest ellipse holding a lattice point has area at most four times the determinant ×1
- the starting arrow has a length ×1
- the starting points are between three and twelve pairs of numbers ×1
- the struck row is one of the vertices ×1
- the three columns are not flat, or there is no solid to draw ×1
- the two directions in the circle are perpendicular ×1
- the two stretches are equal, so every direction is stretched alike ×1
- the two stretches multiply to the area factor ×1
- the unsigned sum is at least as large as the signed one ×1
- the view is one the family draws ×1
- the λ window is a positive width up to 20 ×1
- third ×1
- three well-separated eigenvalues ×1
- two different roots give two independent directions ×1
- two equal columns leave no area at all ×1
- two finishing points, on a small grid ×1
- two starting points, on a small grid ×1
- what survives and what is lost add to the source's dimension ×1
- what survives and what is lost add to two ×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.
A determinant that counts trees
Write down a graph's Laplacian, strike out one row and its column, take the determinant. The answer is the number of spanning trees — and the minus signs in the determinant are what cancel every subset of edges that is not one.
DynamicsA point that pulls, and a point that pushes
Every crossing of a curve with the diagonal is a value the rule leaves alone. Whether anything ever arrives there is decided by one number — the slope at the crossing — and the picture makes the reason obvious.
AlgebraCounted across and counted down
A rectangular array has a number of independent rows and a number of independent columns. The two are counted in different spaces, from different objects, by computations that share nothing — and they are always the same number, which is why 'rank' is one word.
AlgebraDiscs that fence in the eigenvalues
Draw one disc for each row of a square matrix, centred on the diagonal entry, with a radius equal to the sum of the sizes of everything else in that row. Every eigenvalue lies inside one of the discs, and a group of discs set apart from the rest holds exactly as many eigenvalues as it has discs. Nothing is solved to find them.
AlgebraA matrix is a picture of what happens to the grid
Four numbers in a box is not an object anyone has intuitions about. The same four numbers, shown as an instruction for redrawing the plane, are.
AlgebraOne point in every big enough shape
A determinant measures a lattice, not the basis that happened to describe it — and that measurement is an exchange rate. Any symmetric convex region with more than four times that area has to swallow a lattice point.
GeometryOne sign decides which curve
The general quadratic in two variables has six coefficients and draws a conic. Which of the four it draws is settled by a single combination of three of them, and the other three cannot change the answer however they are chosen.
AlgebraOne subtraction clears a direction
A basis is a set of directions to measure along, and most bases are awkward because the directions get in each other's way. Removing one shadow at a time turns any basis into one where every coordinate is a shadow and nothing interferes.
TopologySomething always stays put
Stir a cup of coffee however violently and let it settle. Some molecule is exactly where it started. Crumple a map and drop it on the region it depicts, and one point lies over the place it names.
AlgebraSymmetry forces a right angle
A matrix equal to its own reflection across the diagonal always has real stretches and always has perpendicular directions to stretch along. Neither is true of matrices in general, and both follow from one line of algebra.
AlgebraThe cycles and the cuts
The count that splits a map's source into what dies and what survives has nothing to do with graphs. Apply it to a matrix built from a graph's edges and points and it says that a graph's independent cycles and its independent cuts add to its number of edges — a theorem about drawings, obtained from an array.
AlgebraThe directions a map leaves alone
Almost every arrow comes out of a transformation pointing somewhere else. A few come out pointing exactly where they went in, only longer or shorter. Those few decide nearly everything the map does.
AlgebraThe dot product is a shadow
Multiply the matching coordinates and add them up. That rule explains nothing, and it hides the fact that the answer is a length — how far one arrow reaches along another, times how long that other one is.
AnalysisThe exponential of a square
The series for e makes perfect sense with a matrix in it. What comes out solves a system of equations the way the ordinary exponential solves one, and a skew matrix exponentiates into a rotation with no trigonometry anywhere.
AnalysisThe flat map that fits closest
A derivative is usually met as a number, which works because a line through a point is described by one. In more than one dimension the object that plays the same role is a linear map, and the number was always a one-by-one instance of it.
AlgebraThe highest point on the sphere is an eigenvalue
For a symmetric matrix, walk a unit arrow over every direction and record the value of xᵀAx. The highest value reached is the largest eigenvalue, the lowest is the smallest, and every eigenvalue in between is a saddle height, a minimum of maxima. From that one description comes a theorem no formula for the roots could give: delete a row and its column, and every eigenvalue of what is left sits between two of the original's.
AlgebraThe nearest point of a flat thing
More equations than unknowns almost never have a solution. Asking instead for the point of a plane nearest to where the answer should have been turns an unanswerable question into a shadow, and the shadow is what a line of best fit is.
AlgebraThe number that says how much room is left
A linear map takes the unit square to a parallelogram. The area of that parallelogram is one number, it is computable from the four entries of the matrix, and almost everything the determinant is used for is a restatement of that sentence.
AlgebraThe only function that behaves like a volume
Ask for a function of the columns of a matrix that scales when a column scales, vanishes when two columns agree, and gives one on the identity. Three conditions, and there is exactly one such function in every dimension.
AlgebraThe polynomial whose roots are the stretches
Finding the directions a map leaves alone means finding the numbers at which it crushes something to nothing. Those numbers are the roots of one quadratic, and everything the map does is written in its two coefficients.
AlgebraThe same map in a better basis
Measured along its own invariant directions, a linear map stops shearing and becomes two independent stretches. Nothing about the map has changed; the grid it is described against has.
AlgebraThe same sum without its minus signs
Delete the signs from the determinant's sum over permutations and what is left counts things directly rather than by cancellation. It is a better count and a far worse object — because the cancellation was what made the determinant computable.
AnalysisThe slope of the mirror image
Undoing a function is reflecting its graph in the diagonal, and a reflection turns a slope into its reciprocal. That single observation supplies the derivative of every inverse — the logarithm, the roots, the inverse trigonometric functions — without differentiating any of them.
GeometryTwo right angles and the diagonal of a box
The theorem applied once gives the diagonal of a floor. Applied again, standing on the first result, it gives the diagonal of the room — and the pattern does not stop at three, which is where a fact about triangles quietly becomes the definition of distance.
AlgebraWhat a map does to a circle
Every linear map sends the unit circle to an ellipse. Two perpendicular directions go to two perpendicular directions, whatever the map is — even a map with no invariant direction at all, and even one that is not square.
AlgebraWhat a map throws away
A linear map redraws the grid, and the determinant measures how much it stretches area. When that measurement comes out zero the map has flattened the plane onto a line — and the question worth asking is not how much was lost but how much survived, because the two always add to what there was.