Bourbaki

§ 2. Graphs

1. Definition of a graph

Definition 1 01SU

A graph[^1] is a quiver (S,F,o,t)(S,F, o, t) endowed with an involution of F, denoted by f↦f‾f\mapsto \overline{f}, without fixed point and such that t(f‾)=o(f)t(\overline{f}) =o(f) for every f∈Ff\in F.

The quiver (S,F,o,t)(S,F, o, t) is said to underlie this graph. For every arrow ff of this quiver, one has, by definition, f‾=f,f‾=f̸\overline{f}=f,\overline{f}=\not f and t(f‾)=t(\overline{f}) = o(f)o(f). Applying this last relation to the arrow f‾\overline{f}, one obtains the equality o(f‾)=t(f)o(\overline{f}) =t(f). The arrow f‾\overline{f} is called the arrow opposite to ff.

A pair of opposite arrows of the graph is called an edge of the graph. Each of the two arrows belonging to this pair is called an orientation of this edge. For this reason, the arrows of a graph are also called the oriented edges of the graph.

If G and G′G' are graphs, a graph morphism of G into G′G' is a quiver morphism φ:G→G′\varphi : G\rightarrow G' such that φ(f)=φ(f‾)\varphi (f) =\varphi (\overline{f}) for every arrow ff of G.

Let G and G′G' be graphs; one says that G′G' is a subgraph of G if it is a subquiver thereof and if the involution of Fl(G′') is the restriction of that of Fl(G).

2. Orientation of a graph

Let G be a graph. An orientation of G is a subset A of the set of arrows of G such that A∩A‾=∅A\cap \overline{A}=\emptyset and A∪A‾=A\cup \overline{A}= Fl(G).

A graph endowed with an orientation is called an oriented graph.

Let G be an oriented graph and let A be its orientation; an oriented subgraph of G is a subgraph G′G' of G endowed with the orientation A′=A'= Fl(G′')∩A\cap A.

Let (G,A)(G,A) and (G′,A′)(G',A') be oriented graphs. An oriented graph morphism of (G,A)(G,A) into (G′,A′)(G',A') is a graph morphism φ:G→G′\varphi : G\rightarrow G' such that Fl(φ\varphi )(A)⊂A′(A)\subset A'.

3. Oriented graphs and quivers

Let G be a graph, (S,F,o,t)(S,F, o, t) the underlying quiver, and A an orientation of G. Then (S,A,o∣A,t∣A)(S,A, o|A, t|A) is a quiver, called the quiver associated with the oriented graph (G,A)(G,A).

Conversely, let C=(S,F,o,t)C = (S,F, o, t) be a quiver. Put F~=F×{−1,1}\widetilde{F} = F\times \{-1,1\} and let o~,t~\widetilde{o},\widetilde{t} be the mappings of F~\widetilde{F} into S defined by

o~(f,1)=o(f)o~(f,−1)=t(f) \widetilde{o}(f,1) =o(f)\widetilde{o}(f,-1) =t(f)
t~(f,1)=t(f)t~(f,−1)=o(f) \widetilde{t}(f,1) =t(f)\widetilde{t}(f,-1) =o(f)

for f∈Ff\in F. Then C~=(S,F~,o~,t~)\widetilde{C} = (S,\widetilde{F},\widetilde{o},\widetilde{t}), endowed with the involution (f,ε)↦(f,−ε)(f, \varepsilon )\mapsto (f,-\varepsilon ) of F~\widetilde{F}, is a graph, called the graph associated with the quiver C. The set A=F×{1}A = F\times \{1\} is an orientation of this graph. One says that (C~,A)(\widetilde{C},A) is the oriented graph associated with the quiver C.

If ff is an arrow of C, one also says that (f,1)(f,1) is the oriented edge of C~\widetilde{C} associated with ff and that the pair {(f,1),(f,−1)}\{(f,1),(f,-1)\} is the edge of C~\widetilde{C} associated with ff.

If C′C' is a subquiver of C, the oriented graph C~′\widetilde{C}' is an oriented subgraph of C~\widetilde{C}.

There exists a unique quiver morphism φ\varphi from C into the underlying quiver of C~\widetilde{C} such that φ(f)=(f,1)\varphi (f) = (f,1) for every arrow ff of G; it is an isomorphism of C onto the quiver associated with the directed graph (C~,A)(\widetilde{C},A).

4. Trees

When one speaks of paths, or connected components, of a graph, one means the paths or connected components of the underlying quiver. Two vertices of a graph belong to the same connected component if and only if there exists a path joining them.

Let G be a graph. If c=(a0,f1,a1,...,fn,an)c= (a_0, f_1, a_1, . . . , f_n, a_n) is a path in G, the sequence c‾=(an,fn,...,a1,f1,a0)\overline{c}= (a_n, f_n, . . . , a_1, f_1, a_0) is a path in G, called the opposite path to cc. Let cc and c′c' be juxtaposed paths in G; then c′‾\overline{c'} and c‾\overline{c} are juxtaposed, and one has c∗c′‾=c′‾∗c‾\overline{c*c'}=\overline{c'}*\overline{c}.

A path cc in G is said to be without backtracking if there exists no pair of consecutive arrows of cc which are opposite. Let c=(a0,f1,...,fn,an)c= (a_0, f_1, . . . , f_n, a_n) be a path in G joining a0a_0 and ana_n. If, for an integer ii such that 1⩽i<n1\leqslant i < n, the arrows fif_i and fi+1f_{i+1} are opposite, the path (a0,f1,...,ai−1,fi+2,...,fn,an)(a_0, f_1, . . . , a_{i-1}, f_{i+2}, . . . , f_n, a_n) is a path in G joining a0a_0 to ana_n whose length is strictly less than that of the path cc. By induction, there therefore exists a path without backtracking in G which joins a0a_0 to ana_n.

Definition 2 01SV

A forest is a graph in which every loop without backtracking is a constant loop. A tree is a connected forest.

Proposition 1 01SW

Let G be a graph. Every forest of G is contained in a maximal forest of G; in particular, there exists a maximal forest of G.

In order that a forest of G be maximal, it is necessary and sufficient that the set of its vertices be equal to the set of vertices of G and that its connected components be those of G.

Let A0A_0 be a forest of G. The set of forests of G, endowed with the order relation “A is a subgraph of B”, is inductive. Consequently there exists a maximal forest of G of which A0A_0 is a subgraph (E, III, p. 21, Cor. 1).

The subgraph of G whose set of vertices is Som(G) and whose set of arrows is empty is a forest of G. There therefore exists a maximal forest of G.

Let A be a maximal forest of G. Let us prove that A and G have the same set of vertices. The subgraph of G whose set of vertices is Som(G) and whose set of arrows is Fl(A) is a forest and has A as a subgraph. Hence Som(A) = Som(G).

Let us now prove that A and G have the same connected components. Since an arrow of A is an arrow of G, every connected component of A is contained in a connected component of G. Since A and G have the same set of vertices, it is sufficient to prove that two vertices of G which are in the same connected component of G are in the same connected component of A. If this is not the case, the relation RAR_A, “being in the same connected component of A”, is strictly finer than the relation RGR_G, and there exist two vertices of G which are not in the same connected component of A but which are nevertheless joined by an arrow ff of G. Let B be the oriented subgraph of G whose set of vertices is Som(G) and whose set of arrows is Fl(A) ∪{f,f}\cup \{f, f\}; let us prove that B is a forest of G. Let c=(a0,f1,...,fn,an)c= (a_0, f_1, . . . , f_n, a_n) be a non-constant loop without backtracking in B of minimal length. Since A is a forest, the loop cc is not a loop in A. Let ii (resp. jj) be the smallest (resp. the largest) integer of {0,...,n}\{0, . . . , n\} such that a0a_0 and aia_i (resp. aja_j and ana_n) are not in the same connected component of A. This means that the arrows fif_i and fj+1f_{j+1} are oriented arrows of B associated with the edge {f,f}\{f, f\} and that they are opposite. Since the loop cc is without backtracking, fi+1≠fi‾f_{i+1}=\not\overline{f_i}, hence i=j̸i=\not j and the path (ai,fi+1,ai+1,...,fj,aj)(a_i, f_{i+1}, a_{i+1}, . . . , f_j, a_j) is a non-constant loop without backtracking in B of length <n< n, contrary to the assumption that cc is of minimal length. It follows that B is a forest. This contradicts the assumption that A is a maximal forest of G.

Let now A be a forest of G such that Som(A) = Som(G) and π0(A)=π0(G)\pi_0(A) =\pi_0(G); let us prove that it is a maximal forest of G. It is enough to prove that, if f∉f\notin Fl(A), the subgraph B of G with set of vertices Som(G) and set of arrows Fl(A) ∪{f,f}\cup \{f, f\} is not a forest. By hypothesis, the points o(f)o(f) and t(f)t(f) are in the same connected component of A; there therefore exists a path cc without backtracking in A joining o(f)o(f) to t(f)t(f). Then the path c∗f‾c*\overline{f} is a nonconstant loop without backtracking in B, which shows that B is not an oriented forest.

Corollary 01SX

A maximal forest of a connected graph is a maximal tree in it.

Remark 1 01SY

In LIE, IV, p. 33, appendix, the notion of a combinatorial graph was defined as a pair (A,S)(A,S), where S is a set and A a subset of P(S)\mathfrak{P}(S) consisting of sets with two elements; the elements of S are called vertices, those of A edges, and two vertices xx and y∈Sy\in S are said to be joined if {x,y}\{x, y\} is an edge.

To such a combinatorial graph Γ=(A,S)\Gamma = (A,S), one associates a graph G whose set of vertices is S and whose set of arrows A~\widetilde{A} is the subset of S2S^2 consisting of pairs of joined vertices, the origin mapping and the end mapping coinciding with the first and second projection of S2S^2 onto S, and the involution f↦f‾f\mapsto \overline{f} being given by the restriction to A~\widetilde{A} of the mapping (x,y)↦(y,x)(x, y)\mapsto (y, x) of S2S^2 into itself. The mapping which associates with an edge {f,f}\{f, f\} of G the set {o(f),t(f)}\{o(f), t(f)\} is a bijection of the set of edges of G onto that of the edges of the combinatorial graph Γ.

Conversely, every graph such that the origin and the end of every arrow are distinct, and such that an arrow is determined by its origin and its end, is of this form.

The reader will verify that the notions of connectedness, tree, and forest for a combinatorial graph coincide with the corresponding notions for the graph associated with it.

Exercises

See the exercises for § 2.

[^1]: Care will be taken not to confuse the notion of graph introduced here with that of E, II, §3, no1^o1.

Machine translation of the English, by gpt-5.4, against glossary version 34. Not checked by a person.