Given the difficulty associated with defining a line in taxicab geometry, we won't discuss the parallel postulate. However, we can discuss axioms 6 through 8.

How would we define an isometry of taxicab space?

Is it true that given any two points in taxicab space, we can find an isometry that takes one to the other?

Given a point P and two points Q and R so that |PQ| = |QR|, is it true that there's an isometry of taxicab space that fixes P and moves Q to R?

Are there any reflections that are isometries of taxicab space? If so, what are they?

Using linear algebra, we can describe all of the isometries of taxicab space that fix a specified point.

Think of the point that stays fixed as the origin, (0,0). Let your x-axis run east/west from the origin, and your y-axis run north/south. Then you can label the intersections in your street grid with Cartesian coordinates: (0,1) is one block north of (0,0); (1,0) is one block east of (0,0); and so on.

If an isometry has to leave (0,0) where it is, what can it do with the rest of the points in taxicab space? It can move (1,0) to: (1,0), (0,1), (-1,0) or (0,-1). Any other point is too far away from (0,0); isometries have to preserve distance.

It can move (0,1) to (1,0), (0,1), (-1,0) or (0,-1), but whichever of those it moves to it must be the next point either clockwise or counterclockwise from the image of (1,0).

Linear algebra lets us represent all of these possibilites using matrices. To find out where the isometry corresponding to a matrix sends a point, we multiply on the left by a 2 by 2 matrix. The matrix shown below fixes all the points whose x and y coordinates are equal, and swaps the x and y coordinates of all the other points in taxicab space.

[0 1] [a] [b] [ ]*[ ] = [ ] [1 0] [b] [a]What are the other isometries that fix (0,0) and preserve taxicab distance? Take some time and make a list of them.

We could use vector addition to describe some isometries that do not fix (0,0). Can we describe all the isometries of taxicab space as combinations of some isometry that fixes (0,0) and a vector addition?

[1 0] [1 0] [0 1] [ 0 1] [-1 0] [-1 0] [0 -1] [0 -1] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [0 1] [0 -1] [1 0] [-1 0] [0 1] [0 -1] [1 0] [-1 0]

In taxicab geometry, there are many pairs of points that do not have any points between them. Taxicab geometry is discrete while Euclidean, hyperbolic and spherical geometry is continuous.

However, the points in the taxicab plane go on forever. In a finite geometry there are only finitely many points!

A very simple example of a finite geometry consists of the four vertices of a square. Each pair of vertices determines a line. In this geometry there are four points and six lines.

What happens to our eight axioms in this situation? It may or may not make sense to talk about isometries; axioms 6 through 8 can often be handled by listing all possible isometries!

With only four points, we're not going to talk much about distance or care much about circles.
Again, we could list all the possible circles if we wanted to, and usually we don't want to.
Angles are similarly uninteresting, and there's no place to extend lines to. In the end, only
the first, second and fifth axioms turn out to be interesting. We also add an axiom to make sure
that there are *some* points in our geometry.

**Axiom F1:** Any two points determine a line. (Notice that this is more a statement
about what lines exist than about what they look like!)

**Axiom F3:** There exist four points, no three of which belong to the same line. (This
forces our geometry to be two dimensional and non-empty.)

**Axiom F2a:** Given a line L and a point P not on L, there exists a unique line through P that
doesn't intersect L.

Notice that combining axioms F3 and F2p forces a certain amount of structure on your geometry. If we try to declare that our geometry will have five points A, B, C, D and E and that any three consecutive letters determine a line, we can't find a line through D parallel to the line containing A, B and C. Finding definitions of points and lines that satisfy these axioms is tricky but useful.

**Axiom F2b:** Any two distinct lines intersect in exactly one point.

Finite geometries that obey axioms F1 and F3 and this version of axiom F2 are also tricky to find and useful to have. The smallest and most famous of these geometries is the Fano plane, which consists of seven points and seven lines.

You may remember modular arithmetic from a previous course or from abstract algebra. When
calculating modulo **n** you add, subtract and multiply as usual, but instead of the usual
answer your result is the remainder when the result is divided by **n**. So, when working
modulo 5, 3 + 3 = 1.

If **n** is prime or a power of a prime you can also define division modulo **n**, and
the numbers modulo
**n** form a field. If we put these numbers on the x- and y- axes of a coordinate system,
we'll get **n**^{2} points which form a discrete geometry that obeys the parallel
postulate (Axiom F2a). A similar process can be used to describe a discrete geometry which
obeys Axiom F2b.

With so many examples of finite geometries to study, mathematicians still do not know if there are any finite geometries that behave significantly differently from these examples! (See this Wikipedia article for more information.)