Table des matières

Table des matières i

1 Affine Geometry 1
1.1 Introduction : Parallel Postulate . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Affine spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Definition and examples . . . . . . . . . . . . . . . . . . . . . . 2
1.2.2 Affine subspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Barycenters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4 Affine mappings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Chapitre 1

Affine Geometry

1.1 Introduction : Parallel Postulate

Affine geometry is the geometry of affine spaces : it roughly concerns sets of points
defined by specific properties that allow for discussions of alignment, parallelism, and
intersection. However, the notions of length and angle are foreign to it ; these depend
on additional structures treated within the framework of Euclidean geometry.

The separation of concepts specific to affine geometry is relatively recent in the

history of mathematics. The current formal definition of an affine space presupposes
the existence of a vector space, referred to as the directing space. Two points in an
affine space can be subtracted to yield a vector from the directing space.

Among the remarkable results of affine geometry, we can mention :Thales’ theorem,
the associativity of the centroid, Menelaus’ theorem, Ceva’s theorem, etc.

There are the axioms of Euclid, which establish relationships between objects called
points and others called line.

Postulate 1 : Through two points passes a straight line and only one.
Points lie on lines, and lines intersect at points. For what will concern us here, two
lines that do not meet (or are identical) are said to be parallel (denoted as DkD0 ). One
of Euclid’s axioms, the famous "fifth postulate," can be stated as follows (this is not
Euclid’s original formulation, but it is equivalent to it) :
Chapitre 1. Affine Geometry 2

Postulate 2 : Through a point not on a line, there passes a unique line parallel to
that line in the same plane.

This fifth postulate has the following consequence :

Proposition 1.1.1. The relation "being parallel to" is an equivalence relation between
lines in the plane.

Proof : By definition, it is reflexive and symmetric. Let us show that it is transitive.

We assume that three lines D, D0 , and D00 are such that DkD0 and D0 kD00 . We want to
show that D and D00 are parallel. Suppose D ∩ D00 is not empty, let A ∈ D ∩ D00 . The
line D0 passes through A and is parallel to both D and D00 . Therefore, we have D = D00 ,
thanks to the uniqueness in the fifth postulate.

1.2 Affine spaces

1.2.1 Definition and examples

Definition 1.2.1. A set E is endowed with the structure of an affine space by the data
of a vector space E and a mapping ϕ that associates a vector of E with any ordered
pair of points in E,
ϕ:E ×E −→ E
−→
(A, B) 7−→ AB
such that :
−→ −→ −−→
(i) for all points A, B and C in E, we have AB = AC + CB (Chasles relation) ;
−→
(ii) for any point A of E, the partial map ϕA : B 7−→ AB is a bijection from E to
E.
→
−
The vector space E (also noted E ) is the direction of E, or its underlying vector space,
the elements of E are called points. The dimension of the vector space E is called the
dimension of E.

Example 1.1. 1. Any vector space has the natural structure of an affine space : the
mapping
ϕ:E×E →E
is simply the mapping that associates with the ordered pair (u, v) the vector v − u. 2.

Proposition 1.2.1. Let E be an affine space.

Chapitre 1. Affine Geometry 3

−→
1. For all A ∈ E, AA = ~0 ;
−→ −→
2. For all A, B ∈ E, BA = −AB ;
−→ −→
3. For all A, B, C ∈ E, if AB = AC, then B = C ;
−→
4. For all A, B ∈ E, if AB = ~0, then A = B ;
−→
5. For all A, B ∈ E, there exists a unique →
−
u ∈ E such that →
−
u = AB ;
6. For all A, B, C, D ∈ E, the following properties are equivalent :
−→ −−→
(i) AB = CD ;
−→ −−→
(ii) AC = BD ;
−→ −→ −−→
(iii) AB + AC = AD.
If one of these conditions is met, we say that A, B, C, D form (in this order)
the parallelogram ABCD.

Proof.
−→ −→ −→ −→
1. From axiom (i), we have : AA + AA = AA, hence AA = ~0 (vector calculation
rule).
−→ −→ −→ −→ −→
2. Again using (ii), we have : AB + BA = AA, so BA = −AB by 1).
−→ −→
3. We have : AB = AC implies ϕA (B) = ϕA (C), and since ϕA is bijective (ii), it is
injective, so B = C.
−→ −→ −→
4. Using 1) and 3), we have : AB = ~0 implies AB = AA, hence B = A.
5. This is evident (obvious), it is the translation of the fact that ϕ is an application.
6. Exercise.

Proposition 1.2.2. Let E, F be two affine spaces. Then the Cartesian product E × F
is naturally equipped with a structure of an affine space associated to E × F , and we
have dim(E × F) = dim E + dim F.

Proof. It is easy to see that the map

→
− → −
ϕ : (E × F) × (E × F) → E ×F
−→ −−→
((A, A0 ), (B, B 0 )) 7 → (AB, A0 B 0 )

satisfies the two axioms defining an affine space.

The existence of vector spaces in every dimension having been proved in the linear
algebra course, we conclude :

Corollary 1.6. There exist affine spaces in every dimension.

Chapitre 1. Affine Geometry 4

Remark.
1. If A is a point of the affine space E and if ~u is a vector of the vector space E, the
−→
unique point B of E such that AB = ~u is sometimes denoted B = A + ~u, is the notation
of Grassmann (this does not correspond to a usual additive internal law).
2. For ~u ∈ E, the mapping T~u : E → E which associates to each point A ∈ E the point
A + ~u is a bijection from E onto E, called the translation by the vector ~u.
3. For A ∈ E, the mapping ϕA : E → E which associates to each point M ∈ E the vec-
−−→
tor AM is a bijection from E onto E. Its inverse bijection is the mapping ψA : E → E
defined by ψA (~u) = A + ~u for all ~u ∈ E.

By using this remark, we can also give a second definition of affine space equivalent
to first one.

Definition 1.2.2. An affine space of direction E is a set E together with a map θ :

E × E −→ E defined by (A, ~u) 7−→ A + ~u, which satisfies the following properties :
(i) ∀A ∈ E, A + ~0 = A;
(ii) ∀A ∈ E, ∀~u, ~v ∈ E, we have (A + ~u) + ~v = A + (~u + ~v );
(iii) Given A, B ∈ E, there exists a unique ~u ∈ E such that B = A + ~u.

−→
The unique vector ~u is denoted by AB or B − A.
Vectorization of an affine space. Once a point A has been chosen in an affine space
E, it is possible to give E the structure of a vector space. This vector space is denoted
EA . The mapping
ϕA : E −→ E
−−→
M 7−→ AM
is a bijection, and it allows us to transfer the vector space structure from E to E. It
−−→ −−→ −→
is said that M + N = P if AM + AN = AP . Notice that the vector space structure
defined in this way depends heavily on the point A, which becomes the zero vector of
−→
the vector space EA due to the relation AA = ~0.

1.2.2 Affine subspaces

In this entire section, E is an affine space, with vector space E as its direction.

Definition 1.2.3. A non-empty subset F of E is an affine subspace of E if is contains

a point A such that ϕA (F) is a vector subspace of E.
Chapitre 1. Affine Geometry 5

Remark. By the following proposition, we show that ϕA (F) does not depend on
the choice of the point A.

Proposition 1.2.3. Let F be an affine subspace of E. There exists a vector subspace F

of E such that, for any point B of F, ϕB (F) = F . The subspace F is an affine space
directed by F .

It follows that, for A ∈ F, we have

−−→
F = {AM ; M ∈ F}.

Proof. By assumption, there exists A ∈ F such that ϕA (F) is a vector subspace of

−→
E. Fix any B ∈ F and show that ϕA (F) = ϕA (F). We have AB = ϕA (B) ∈ ϕA (F).
−−→ −→
For any M ∈ F, the vectors AM and AB belong to the vector subspace ϕA (F), so
−−→ −−→ −→
BM = AM − AB ∈ ϕA (F). This proves that ϕB (F) ⊆ ϕA (F). Note that we also have
−−→ −→ −−→ −→ −−→
AM + AB ∈ ϕA (F). Therefore, there exists N ∈ F such that AM + AB = AN , hence
−−→ −−→ −→ −−→
AM = AN − AB = BN ∈ ϕB (F). We conclude that ϕA (F) ⊆ ϕB (F), which gives the
desired equality.

Conversely, the data of a vector subspace of E and a point in E uniquely determines

an affine subspace of E, as shown by the following theorem.

Theoreme 1.2.1. Let F be a vector subspace of E and A a point in E. There exists a

unique affine subspace F of E such that A ∈ F and F is the directing vector subspace
of F.
We say that F is the affine subspace of E passing through A and directed by F and
dim F = dim F .

In summary : If F is an affine subspace of E and F is the directing vector subspace

of F, we have the following three properties :
−→
(a) For all A ∈ F and B ∈ F, we have AB ∈ F ;
−−→
(b) For all A ∈ F and ~u ∈ F , there exists a unique point M ∈ F such that ~u = AM ;
−−→ −−→
(c) For all A ∈ F, we have : F = {M ∈ E | AM ∈ F } and F = {AM ; M ∈ F}.

Example 1.2. 1. An affine subspace of dimension 0 is a singleton A consisting of a

single point in E. An affine subspace of dimension 1 is called an affine line. An affine
subspace of dimension 2 is called an affine plane. If dim F = dim E − 1, we say that F
is an affine hyperplane of E.
2. Let E and F be two vector spaces and let f : E → F be a linear mapping. For any
v in the image of f in F , the inverse image f −1 (v) is an affine subspace of E (here
we consider, of course, that E is endowed with its natural affine structure) directed by
ker f .
Chapitre 1. Affine Geometry 6

Proposition 1.2.4. An intersection of affine subspaces is an affine subspace, directed

by the intersection of the directing vector subspaces.

Affine subspace spanned by a subset of E

Definition 1.2.4. Let S be a no-empty subset of E. We call affine subspace spanned

(or generated ) by S, denoted hSi, the smallest affine subspace containing S (in the
sense of inclusion). This affine subspace coincides with the intersection of all the affine
subspaces of E containing S.

Theoreme 1.2.2. Let A0 , A1 , . . . , Ak be distinct finite points in E. The affine subspace

spanned by {A0 , A1 , . . . , Ak } is equal to the affine subspace of E passing through A0
−−−→ −−−→ −−−→
and directed by the vector subspace spanned by the vectors {A0 A1 , A0 A2 , . . . , A0 Ak }. Its
dimension is at most k.

Remark. It is clear that, in this theorem, A0 can be replaced by any of the points Ai .

In summary, and in practice : A point M in E belongs to the affine subspace spanned

by A0 , A1 , . . . , Ap if and only if there exists a k-tuple (α1 , . . . , αk ) ∈ Rk such that
k
−−−→ X −−−→
A0 M = αi A0 Ai .
i=1

Definition 1.2.5. 1- A family of k+1 points in an affine space E is affinely independent

if its generated affine subspace of the dimension k.
2- A Cartesian frame of an affine space E of dimension n is a set R = {O; (e1 , . . . , en )}
formed by a point O ∈ E and a basis (e1 , . . . , en ) of the associated vector space E. The
point O is called the origin of this affine frame.

Proposition 1.2.5. Let R = {P ; (e1 , . . . , en )} be a Cartesian frame in an affine space

E. For all M ∈ E there exist scalars α1 , . . . , αn such that M = P + ni=1 αi ei . The
P

family (α1 , . . . , αn ) is called Cartesian coordinates of M in this frame.

Proof.
−−→
It suffices to notice that the vector P M belongs to the vector space E whose base is
(e1 , . . . , en ) then there exits αi such that
n
−−→ X
PM = αi ei .
i=1
Chapitre 1. Affine Geometry 7

Definition 1.2.6. If the k +1 points A0 , . . . , Ak are affinely independent and if dim E =

k, we say that {A0 , . . . , Ak } is an affine frame of E

Remark. An affine frame {A0 , . . . , Ak } of E can be considered as the datum of the

−−−→ −−−→
origin A0 and a basis (A0 A1 , . . . , A0 Ak ) of its direction.

Example 1.3. 1- Two points A and B distinct generate an affine line hA, Bi and
−→
(A, AB) is a Cartesian frame.
2- If A and B are two distinct points, the segment AB is the set of points M of the line
−−→ −→
hA, Bi such that AM = λAB with 0 ≤ λ ≤ 1.
3- Three points A, B and C distinct and not aligned generate an affine plane hA, B, Ci
−→ −→ −→ −→
and (A, AB, AC) is a Cartesian frame where {AB, AC} are linearly independent.

Parallelism.

→
−
Definition 1.2.7. Let E be an affine space of direction E and F, G two affine subspaces
→
− →
−
of E directed respectively by F and G .
→
− →
−
i) We say that F and G are parallel, noted FkG, if F = G .
→
− →
−
ii) We say that F is weakly parallel to G, if F ⊂ G .

Remark.
1- The relation to be parallel is an equivalent relation on the family of the affine
subspaces of E. The relation weakly parallel is a partial order relation.
2- Two affine subspaces can be disjoint with out being parallel, for example, a line is
not parallel to a plane.

Example 1.4. If f : E −→ F is a linear mapping between to vectors spaces, all the

affine subspace f −1 (v) where v ∈ Imf are parallel since they have the same direction
ker f .

Exercise 1. Let A, B, A0 , B 0 be four distinct points in an affine plane, with no three

of them collinear. Show that :
−→ −−→
(AB)||(A0 B 0 ) and (AA0 )||(BB 0 ) ⇐⇒ AB = A0 B 0 .
→
−
Proposition 1.2.6. Let F, G two affine subspaces of E directed respectively by F and
→
−
G.
Chapitre 1. Affine Geometry 8

— If F and G are parallel then F and G are disjoint or equal.

— If F is weakly parallel to G then F ⊂ G or F ∩ G = .

Now, we can proof the parallel axiom in the affine spaces.

Proposition 1.2.7. Let F and G be two affine subspaces of an affine space E, directed
respectively by F and G. Let A be a point of F, B a point of G. The subspace F ∩ G is
−→
not empty if and only if the vector AB belongs to F + G. In particular : if F and G are
supplementary in E (E = F ⊕ G) then F ∩ G consists of one point.

1.3 Barycenters

→
−
Let E be an affine space directed by a K−vector space E = E .
Definition 1.3.1. Let A1 , . . . , Ak be a distinct points on an affine space E and α1 , . . . , αk
are scalars, we say that the system {(A1 , α1 ), . . . , (Ak , αk )} is a system of punctual
masses (or a system of weighted points).
Proposition 1.3.1. Let ((Ai , αi ))ki=1 is a system of punctual masses such that ki=1 αi =
P

1, there exists a unique point G of E satisfying one of the following equivalent condi-
tions :
−−→ P −−→
(a) ∀M ∈ E : M G = ki=1 αi M Ai ;
−−→
(b) ki=1 αi GAi = ~0.
P

The unique point G is called the barycenter of the system.

Remark.
If α1 , . . . , αk are scalars such that ki=1 αi = α 6= 1(6= 0), then the barycenter of the
P

system of weighted points ((Ai , αi ))ki=1 is the barycenter of system ((Ai , ααi ))ki=1 , noted
bu Bar((Ai , αi )ki=1 ).
Definition 1.3.2. If all the masses αi = 1, i = 1 . . . , k, the barycenter is called equi-
barycenter of the points A1 , . . . , Ak . The equibarycenter of two points A and B is the
midpoint of the segment AB ; that of three noncollinear points A, B and C is called the
centroid or center of gravity of the triangle ABC.
Proposition 1.3.2 (Associativity of the barycenter). Let E be an affine space. Let
((Ai , αi ))i∈I a finite family of of weighted points of E and I = ∪j∈J Ij the partition of
I such that for all j ∈ J, µj = i∈Ij αi 6= 0, α = j∈J µj = i∈I αi 6= 0. If for all
P P P

j ∈ J, Gj is the barycenter of the system ((Ai , αi ))i∈Ij , then we have :

Bar((Ai , αi )i∈I ) = Bar((Gj , µj )j∈J ).

Chapitre 1. Affine Geometry 9

Remark.
In other words, to find the barycenter of a large system, we can first group the terms
and consider their barycenters, then take the barycenter of the system consisting of
these new points, weighted by suitable coefficients (sums of the masses of the points
used in the grouping). A simple and useful special case gives the following corollary :

Corollary 1.3.1. Let G be the barycenter of ((A, α1 ), (B, α2 ), (C, α3 )), where α1 + α2 +
α3 6= 0 and α2 + α3 6= 0. The intersection point I of the lines hA, Gi and hB, Ci is the
barycenter of ((B, α2 ), (C, α3 )).

As a consequence, in a triangle, the three medians (lines from the vertices to the
midpoints of the opposite sides) are concurrent at the centroid of the triangle.

Theoreme 1.3.1 (Effect on affine subspaces). Let F be a non-empty set of an affine

space E. F is an affine subspace of E if and only if for all family (Ai )i∈I (I finite) and
all scalars αi , i ∈ I such that i∈I αi 6= 0, the barycenter of the system ((Ai , αi ))i∈I is
P

in F.

Corollary 1.3.2. Let F be a non-empty set of an affine space E. The following asser-
tions are equivalent :
i) F is an affine subspace of E ;
ii) ∀A, B ∈ F, ∀λ ∈ R we have Bar((A, λ), (B, 1 − λ)) ∈ F,
iii) ∀A, B ∈ F, the affine line hA, Bi is include in F.

Proof. This is a consequence of the previous theorem and the associativity of the
barycenter.

Proposition 1.3.3 (Barycenter coordinates). Let (A0 , . . . , An ) be an affine frame in E

(dim E = n). For all M ∈ E there exists an unique family of scalars (λ0 , . . . , λn ) such
that n
M = Bar((Ai , λi ))ni=0 .
X
λi = 1 and
i=0

The scalars (λ0 , . . . , λn ) are called the barycenter coordinates of M in the barycenter
affine frame (A0 , . . . , An ) of E and we write :
n
X
M= λi Ai .
i=0
Chapitre 1. Affine Geometry 10

1.4 Affine mappings

Definition 1.4.1. A map f : E −→ F between two affine spaces is an affine map if

and only if it preserve barycenter more precisely : if ((Ai , αi ))i∈I is a weighted point
system of E such that i∈I αi 6= 0 and G its barycenter, then f (G) is the barycenter of
P

the system ((f (Ai ), αi ))i∈I .

Remark.
1- By associativity of the barycenter, it suffices to verify this property for any system
of two points.
2- Any constant map is affine, as well as the Identity map.

Example 1.5. Let E an affine space and ~u ∈ E. The translation map T~u : E −→ E
defined by T~u (M ) = M + ~u is an affine map.

Lemma 1.4.1. Let f : E −→ F be an affine map. For any point A of E, the map
−−−−−−−−−→
φA : E −→ F defined by φA (~u) = f (A)f (A + ~u) is linear and satisfies, for all M and
N in E :
−−−−−−−→ −−→
f (M )f (N ) = φA (M N ).
Furthermore, for any point B of E, we have φA = φB .

Lemma 1.4.2. Let E and F be two affine spaces, φ : E −→ F a linear map and A
(resp. B) a point of E (resp. F). Then, the map f : E −→ F defined, for all M ∈ E,
−−→
by f (M ) = B + φ(AM ) is affine.

These two lemmas prove the following theorem :

Theoreme 1.4.1. Let E and F be two affine spaces. A map f : E −→ F is affine if,
and only if there exists a linear amap φ : E −→ F such that for any pair (M, N ) of
points of E, we have :
−−−−−−−→ −−→ −−→
f (M )f (N ) = φ(M N ) i.e. f (N ) = f (M ) + φ(M N ).

In this case, such a linear map is unique and is called a linear map associated with f ,
denoted f~.

Proposition 1.4.1. Two affine maps f, g : E −→ F coincide if, and only if, the two
following conditions are verified :
1. there exists a point A of E such that f (A) = g(A) ;
2. f~ = ~g .
Chapitre 1. Affine Geometry 11

Proof.
The two conditions are obviously necessary. If they are satisfied, for every point M ∈ E
we have :
−−→ −−→
f (M ) = f (A) + f~(AM ) = g(A) + ~g (AM ) = g(A).

Proposition 1.4.2. 1. The linear map associated to IdE is Id− ~ E = Id−

→ : Id →.
E E
2. Let f : E −→ F and g : F −→ G two affine map. Then, g ◦ f is an affine map
−−→
and g ◦ f = ~g ◦ f~.
3. An affine map f : E −→ F is a bijection if, and only if, f~ : E −→ F is an
isomorphism. In this case, f −1 is affine and
−→
f −1 = (f~)−1 .

Proof. Exercise.

Corollary 1.4.1. The affine bijections from E to itself (called an affine automorphism
of E) form a group for the composition, called affine group of E, and denoted, GA(E).

Proposition 1.4.3 (Fixed points of an affine map). The set of fixed points of an affine
map f of an affine space E in itself, noted F ix(f ) = {M ∈ E | f (M ) = M } is either
empty, or an affine subspace of E with direction ker(f~ − IdE~ ).

Theoreme 1.4.2. An affine map f of an affine space E into itself admits a unique
fixed point if, and only if, 1 is not eigenvalue of f~.

Proof.

Proposition 1.4.4 (Effect on the affine subspaces). Let f : E → F be an affine map

with associated linear map φ : E → F .
(i) Let H be an affine subspace of E directed by a vector subspace H of E. Then
f (H) is an affine subspace of F, directed by the vector subspace φ(H) of F .
(ii) Let H0 be an affine subspace of F directed by a vector subspace H 0 of F . If
f −1 (H0 ) 6= ∅, then f −1 (H0 ) is an affine subspace of E, directed by the vector
subspace φ−1 (H 0 ) of E.

Theoreme 1.4.3. Let (A0 , . . . , An ) be an affine frame of an affine space E and (B0 , . . . , Bn )
n + 1 points of an affine space F. There exists a unique affine map f from E to F such
that f (Ai ) = Bi for all i = 1, . . . , n.
Chapitre 1. Affine Geometry 12

Translations and homotheties

→
− →
−
Definition 1.4.2. Let E be an affine space directed by E and ~u ∈ E . The map

T~u : E → E
M 7→ T~u = M + ~u
is called the translation by the vector ~u. In the words :
−−−→0
M 0 = T~u (M ) ⇐⇒ M M = ~u.

It follows from this definition that : τ~0 = idE , and if ~u 6= ~0, then τ~u has no fixed points.
The translation T~u is a bijection with inverse T−~u .

Theoreme 1.4.4. An affine map f of an affine space E into itself is a translation if,
and only if, it satisfies f~ = IdE~ .

Definition 1.4.3. Let E be an affine space, Ω a point of E and λ a non-zero real

number. We call homothety of ratio λ and of center Ω the map hΩ,λ from E to E defined
−−→
by hΩ,λ (M ) = Ω + λΩM for all M ∈ E. In other words :
−−→0 −−→
M 0 = hΩ,λ (M ) ⇐⇒ ΩM = λΩM .

If λ = 1 then hΩ,1 = IdE . If λ = −1, he homothety hΩ,−1 is called the central

symmetry with center Ω ; it associates to every point M the point M 0 such that Ω is
the midpoint of the segment M M 0 .

Proposition 1.4.5. An homothety of ratio λ is an affine bijective map whose inverse

is the homothety of the same center and ratio λ1 .

Theoreme 1.4.5. An affine map h : E −→ E is an homothety distinct from IdE if, and
only if, its linear map is an (vector) homothety with a ratio different from 1.

Projection and symmetry

Linear algebra reminder. Let E be a vector space, F and G two additional vector
subspaces : E = F ⊕ G. Any element ~u of E is then uniquely written in the form

~ with ~v ∈ F, w
~u = ~v + w ~ ∈ G.

The map p : E → E which, to every ~u ∈ E decomposed in this way, associates its

component ~v on F is called the (vector) projection of E onto F parallel to G. We will
Chapitre 1. Affine Geometry 13

note q the projection on G parallel to F .

The map s : E → E which, to every ~u ∈ E decomposed in this way, associates the
vector ~v − w
~ is called the (vector) symmetry with respect to F parallel to G.
It is easy to verify that p, s and q are linear maps, and satisfy

p ◦ p = p, q ◦ q = q, p + q = IdE , s ◦ s = IdE and s = p − q = 2p − IdE .

Lemma 1.4.3. 1. A linear map p : E −→ E is a projection if, and only if, p◦p = p.
In this case, we have E = ker(p − IdE ) ⊕ ker p and p is the projection onto
ker(p − IdE ) = Imp parallel to ker p.
2. A linear map s : E −→ E is a symmetry if, and only if, s ◦ s = IdE . In this
case, E = ker(s − IdE ) ⊕ ker(s + IdE ) and s is the symmetry with respect to
ker(s − IdE ) parallel to ker(s + IdE ).

Theoreme 1.4.6. Let E be an affine space, F an affine subspace of E and G a supple-

mentary space of F ~ in E.
~ Then, for any point M of E, the intersection of F and M + G
is reduced to a point ; if p(M ) is this point, the map p : E −→ E thus defined is affine
with a linear associated map is the vector projection on F ~ parallel to G. The map p is
called (affine) projection on F in the direction G and satisfies p ◦ p = p.

If M ∈ F, then p(M ) = M because M ∈ F ∩ (M + G), i.e., the set of fixed points

of p is F.

Proposition 1.4.6. An affine map p : E −→ E is a projection if, and only if, p ◦ p = p.

Theoreme 1.4.7. Let E be an affine space, F an affine subspace of E and G a sup-

plementary space of F ~ in E.
~ For M ∈ E, we note s(M ) the point defined by s(M ) =
−−−−−→
M + 2M p(M ) where p is the projection on F parallel to G. Then s : E −→ E is an
affine map with a linear associated map is the vector symmetry map with respect to
~ parallel to G. The map s is called (affine) projection on F in the direction G and
F
satisfies s ◦ s = IdE .

From previous theorem, we have for M ∈ E :

1. The point p(M ) is the midpoint of the segment [M ; s(M )],
−−−−−→
2. s(M ) = p(M ) + M p(M ).
3. p(s(M )) = p(M ).

Proposition 1.4.7. An affine map s : E −→ E is a symmetry if, and only if, s◦s = IdE .
Chapitre 1. Affine Geometry 14

Using affine mappings : three theorems in plane geometry

Theoreme 1.4.8 (Thales’ theorem). Let E be an affine space of finite dimension, H a

~ D and D0 two lines of E whose directions are each supplementary
vector hyperplane of E,
~ Let H1 , H2 and H3 be three distinct pairwise hyperplanes of E with the same
to H in E.
direction H, intersecting D (resp. D0 ) at three points M1 , M2 and M3 (resp. M10 , M20
and M30 ). Then
−−−−→ −−− −→
M1 M3 M10 M30
−−−−→ = −−− −→ .
M1 M2 M10 M20
Moreover that, if a point B of D satisfies
−−−→ −−− −→
M1 B M10 M30
−−−−→ = −−− −→ ,
M1 M2 M10 M20

then it is H3 and B = M3 .

−−−−→ −−−−→
Remark. If the points M1 , M2 and M3 are aligned, the vectors M 1 M3 and M1 M2
−−−−→ −−−−→ −−−−→
are collinear : the scalar λ such that M1 M3 = λM1 M2 is noted M 1 M3
−−−−→ = λ.
M1 M2

Corollary 1.4.2. Let D1 and D2 be two lines intersecting at A, and H and H0 two
parallel hyperplanes intersecting Di at Ai , A0i distinct from A. Then
−−→ −−→ −−−→
AA1 AA2 A1 A2
−−→0 = −−→0 = −−0−→0 .
AA1 AA2 A1 A2

Theoreme 1.4.9 (Pappus’ Theorem). Let D and D0 be two distinct lines in the plane E.
Let A, B, C be three points on D and A0 , B 0 , C 0 be three points on D0 . If hA, B 0 i||hB, A0 i
and hB, C 0 i||hC, B 0 i, then hA, C 0 i||hC, A0 i.

Theoreme 1.4.10 (Desargues’ Theorem). Let ABC and A0 B 0 C 0 be two triangles wi-
thout common vertices and whose sides are respectively parallel. Then the lines hA, A0 i,
hB, B 0 i, and hC, C 0 i are either concurrent or parallel.