MATH 130

SPRING, 2012
AFFINE AND PROJECTIVE GEOMETRY
M. J. HOPKINS

1. Affine geometry
Much of this material is taken from Hartshorne [1].
Definition 1.1. An Affine plane is a set A whose elements are called points and
a collection of subsets of A called lines satisfying the properties listed below.
Now slow down there. Before stating the properties we need less fancy terminology. If is a line and P is a point, we will use the colloquialisms contains P
or P lies on and passes through P to indicate that P is an element of the
set . We will also say that a set of points {P1 , . . . , Pn } is collinear if there is a line
containing them all. Finally two lines 1 and 2 are parallel if they are the same
line (1 = 2 ), or if they have no points in common (1 2 = ).
A1. Given two distinct points P and Q there is a unique line containing them
both.
A2. Given a line and a point P , there is one and only one line m which is
parallel to and passes through P .
A3. There exist three non-collinear points.
Before doing anything with this, lets set up some examples. Lets first check
that the thing were most familiar with is one.
Definition 1.2. The Euclidean plane is the affine geometry whose points are pairs
(x, y) of real numbers, and whose lines are the sets of points of the form
a,b,c = {(x, y) | a x + b y + c = 0}
in which at least one of {a, b, c} is non-zero.
Note that the lines a,b,c and a ,b ,c coincide if and only if there is a non-zero
real number with
a = a
b = b
c = c.
We need to check that this data defines an affine plane. Just to give you an
idea, lets prove A1. Once you have the idea, youll be able to deal with the other
axioms. Suppose P = (x1 , y1 ) and Q = (x1 , y2 ) are two distinct points. The fact
that they are distinct means that one of x1 x2 and y1 y2 is non-zero. Suppose
it is x1 x2 . We need to find a line a,b,c containing both P and Q, and we need
to show the line is unique. Expanding out the definition, this means that we need
to solve the equations
ax1 + by1 + c = 0
ax2 + by2 + c = 0
1

M. J. HOPKINS

for a, b, and c, and we need to show that the solution is unique up to multiplication
by . This is a simple problem in linear algebra. You ought to do this by rowreducing. If you do, you get the system of equations
ax1 + by1 + c = 0
a + b(y2 y1 )/(x2 x1 )+ = 0.
Given any b, the second equation determines a, and then the first determines c.
Replacing b by b replaces a by a and c by c. This proves axiom A1. In the
exercise you will handle the other two exercises.
Are there any other examples? Lets try and write down the smallest possible.
Before doing so, it will be helpful to point out a lemma which will be useful later.
Lemma 1.3. In an affine plane, the relation parallel is an equivalence relation.
Proof: The symmetry relation k holds by definition (this is the reason we
allow the case 1 = 2 in the definition of parallel.) The reflexive property is also
immediate from the definition. The real work is in proving that the relation is
transitive. So suppose 1 k 2 and 2 k 3 . If 1 and 3 have no points in common,
they are parallel, and we are done. So suppose they both contain a point P . Then
1 and 3 contain P and are parallel to 2 . By the uniqueness in A2, this means
that 1 = 3 and they are parallel.

On to our tiny plane. We know by A3 that there are at least three points. Lets
call them P , Q, and R. The point R is not on the line P Q, and the point P is not
on QR. By A2 there is a line R containing R and parallel to P Q and a line Q
containing Q and parallel to P R.
(1.4)

PQ

Q
P?
?? 
??
PR
?? Q
 ?
R
S
R

Now Q and R cannot be parallel. If the were we would have

P R k Q k R k P Q
and so P Q and P R would also be parallel by Lemma 1.3. But P R and P Q are not
parallel (why?). This means that there is a point S lying on both Q and R . This
point cannot coincide with P , Q, or R, since each of these points lies on exactly
one of Q or R . Thus S is a unique fourth point. This shows that an affine plane
must contain at least four points. The diagram (1.4) exhibits an affine plane with
exactly four points, and exactly 6 lines.
In the picture, the two lines which appear to cross (P S and QR) are in fact
parallel. In a way it is better to think of the picture as displayed in Figure 1. In
that picture the line QR and P S actually look parallel.
We can try to define an n n affine plane in the same manner. I will leave the
details to you for now. Figure 2 shows the 9-point 3 3 plane, and a few lines in
it. We will see later that the n n plane satisfies the axioms for an affine plane if
and only if n is a prime number.
Note that all the lines in the 4-point plane have exactly 2 points, and all of the
lines in the 9-point affine plane have exactly 3 points. This is not an accident.

AFFINE AND PROJECTIVE GEOMETRY

Figure 1. Two parallel lines in the 4-point plane

Figure 2. Some lines in the 3 3 affine plane

Proposition 1.5. There is a bijection between the points of any two lines in an
affine plane.
Proof: Suppose that 1 and 2 are two distinct lines. Then there is a point P1
on 1 which is not on 2 , and a point P2 on 2 which is not on 1 . Write = P1 P2 .
We may now set up our one-to-one correspondence. Suppose that X is any point
on 1 . By A2 there is a unique line through X parallel to . This line cannot be
parallel to 2 (why?), and so meets 2 in a unique point Y . I leave it to you to
check that this correspondence between points of 1 and points of 2 is a bijection.
See Figure 3 for a picture.

Families of lines come up a lot in affine geometry, and it is useful to have a name
for the kind that do.

M. J. HOPKINS

Figure 3. Bijection between two lines

Definition 1.6. The set of lines through a fixed point p in an affine plane is called
the pencil of lines through p. The set of lines parallel to a fixed line the pencil of
lines parallel to .

Exercises
1.1. Prove A2 for the Euclidean plane as follows. Suppose that a,b,c is a line,
and P = (x, y) is a point not on . Let c = ax by. Show that the point
P lies on the line a,b,c , and that the lines a,b,c and a,b,c are parallel.
1.2. Prove A3 for the Euclidean plane. Make sure you really prove it.
1.3. Let A = R2 \ {0} be the complement of the origin in the Euclidean plane,
and define a line in A to be a subset of the form A, where R2 is a
Euclidean line. Does this define an affine plane?
1.4. Show that the 4 4-plane is not an affine plane.
1.5. Suppose that is a line in an affine plane. Show that there is another line
6= which is parallel to . Show that there is another line 6= which is
not parallel to .
1.6. Show that every line in an affine plane contains at least two points.
1.7. In this exercise we set up coordinate systems. The construction is illustrated
in Figure 4. Suppose that A is an affine plane. Choose a line 1 and another
line 2 which is not parallel to 1 . Given a point P show that the unique
line parallel to 2 and containing P cannot be parallel to 1 . Let x 1 be
the unique point of intersection. Similarly, let y 2 be the unique point
at which the line parallel to 1 and containing P meets 2 . Show that the
correspondence
A 1 2
P 7 (x, y)
is a bijection. Conclude that if an affine plane has one line with n <
points, then it is a finite affine plane and has n2 points.
1.8. Show that an affine plane with n2 points has exactly n2 + n lines.

AFFINE AND PROJECTIVE GEOMETRY

Figure 4. A coordinate system in an affine plane

2. Projective geometry
Things work out a little cleaner if we extend our affine planes by adding ideal
points so that even parallel lines intersect.
Definition 2.1. A projective plane is a set P called the set of points and a
collection of subsets of P called lines satisfying P1-P4 below.
P1.
P2.
P3.
P4.

Given two distinct points, there is a unique line containing them both.
Any two lines meet in at least one point.
There exist three non-collinear points.
Every line contains at least three points.

A little explanation is in order. When P and Q are two distinct points it is common to use the symbol P Q to denote the unique line, guaranteed by P1, containing
them both. As for P2, suppose that 1 and 2 are two lines. By P2 they meet in
at least one point. If they meet in two points, they are the same line by P1. So
P2 could have been written as two distinct lines meet in exactly one point.
Example 2.2. Let A be an affine plane, and a line in A. The ideal point of is
the equivalence class [] of all lines parallel to (or, just to use our new language,
the pencil of lines parallel to ). We define the extended affine plane P to be the
set whose elements are A and the ideal points [] of the lines of A. A line in P
is either an extended line of A, which means a set whose elements consist of the
points of a line and the ideal point [], or the line at infinity which means the set
consisting of all the ideal points []. The ideal point [] is also called the point at
infinity of . With this structure the extended affine plane P is a projective plane.
Two parallel lines 1 and 2 of A now meet at their ideal points, and every line
meets the line at infinity at its ideal point. The projective extension of the 4-point
plane is illustrated in Figure 5. It is sometimes called the Fano plane.
Example 2.3. There is another way to get a projective plane from an affine plane
A. Let P be the set of lines in A, and define a line in P to be a pencil of lines
in A. I leave it to you to check that this data defines a projective plane. I really
do. See the exercises.

M. J. HOPKINS

Figure 5. The four-point plane and its extended (Fano) projective plane
Example 2.4. The Euclidean projective plane or real projective plane, RP2 is the
extended Euclidean plane. It can be described in terms of coordinates as follows.
The set of points of RP2 is the set of all equivalences classes [x, y, z] of triples of
real numbers, at least one of which is non-zero. The equivalence relation is that
[x, y, z] = [x, y, z]
whenever is a non-zero real number. The lines are the sets of the form
a,b,c = {[x, y, z] | ax + by + cz = 0},
in which at least one of {a, b, c} is non-zero. Note that a,b,c = a ,b ,c if and only
if there is a non-zero for which a = a, b = b, and c = c.
The points [x, y, z] with z 6= 0 are called the finite points and the points [x, y, 0]
form the projective line at infinity. This corresponds to the identification of the
real projective plane with the extended Euclidean plane. Indeed, the set of finite
points can be identified with the Euclidean plane by sending [x, y, z] to (x/z, y/z)
and sending (x, y) to [x, y, 1]. The ideal point of a Euclidean line ax + by + c = 0
is [a, b, 0].
One way to think of the real projective plane is in terms of perspective drawing.
Suppose you are sitting at the origin in R3 , and your canvas is the plane z = 1.
When you see a point (x, y, z) you want to draw, you make a mark on your canvas
at the point where the line through (x, y, z) and the origin meets the plane z = 1.
Thus the points on your canvas really correspond to lines through the origin in R3 .
I discussed this point of view further in class.
For our discussion of Desargues theorem, we will need the notion of a projective
3-space.
Definition 2.5. A projective 3-space is a set whose elements are called points,
together with certain subsets called lines, and certain other subsets called planes,
which satisfies the following axioms:
S1 Two distinct points lie on a unique line.
S2 Three non-collinear points lie on a unique plane.
S3 A line meets a plane in at least one point.
S4 Two planes have at least a line in common.
S5 There exist four non-coplanar points, no three of which are collinear.

AFFINE AND PROJECTIVE GEOMETRY

S6 Every line has at least three points.

Example 2.6. The Euclidean projective 3-space RP3 is the set of lines through
the origin in R4 . Its points are equivalences classes [x, y, z, w] of 4-tuples of real
numbers. The equivalence relation is such that [x, y, z, w] = [x, y, z, w], with
6= 0 R. A line is a set of points of the form
T = {[x, y, z, w] | [x, y, z, w] T = 0} ,
where T is a 4 2 matrix with the property that one of its 2 2 sub-matrices has
a non-zero determinant. A plane is a set of points of the form
Pa,b,c,d = {[x, y, z, w] | ax + by + cz + dw = 0}
in which at least one of a, b, c, or d is non-zero. We could write this last condition
as [x, y, z, w] L = 0, with L a non-zero 4 1 matrix. In more geometric terms, a
point of RP3 is a line through the origin in R4 , a line of RP3 is a 2-plane through
the origin in R4 , and a plane of RP3 is a 3-plane through the origin in R4 .
Exercises
2.1. Show that there is a bijection between any two lines in a projective plane.
2.2. Suppose that P is a projective plane. Let P be the set of lines of P, and
define a line of P to be a pencil of lines in P. Show that with this structure,
P is a projective plane.
2.3. Let be a line in a projective plane, and P a point not on . Show that the
map which associates to a point Q of the line P Q gives a bijection between
the points of and the pencil of lines through P .
2.4. Show that if one line of a projective plane P has (n + 1)-points, then P is
finite and has n2 + n+ 1 points. Show that every pencil of lines in P contains
exactly (n + 1) lines. Show that P has n2 + n + 1 lines.
2.5. Suppose that S is a projective 3-space. Show that a line and a plane which
does not contain the line meet in exactly one point, and that two distinct
planes in S meet in exactly one line.
2.6. Suppose that S is a projective 3-space. Let S be the set of planes in S.
Define a line in S to be a pencil of planes through a line of S, and define
a plane of S to be a set of planes in S passing through a given point of S.
Thus there are bijections
points of S planes of S
lines of S lines of S
planes of S points of S.
Show that with this structure S is a projective 3-space.
References
1. Robin Hartshorne, Foundations of projective geometry, Lecture Notes, Harvard University, vol.
1966/67, W. A. Benjamin, Inc., New York, 1967. MR 0222751 (36 #5801)
Department of Mathematics, Harvard University, Cambridge, MA 02138
E-mail address: [email protected]