Bourbaki

§ 1. Quivers

Its monument, superb among monuments,

Which bristles, above a wall of dry bricks, its crest of towers like a quiver of arrows...

Victor Hugo, La Légende des siècles

1. Definition of a quiver

Definition 1 01SS

A quiver is a quadruple (S,F,o,t)(S,F, o, t), where S and F are sets and o,to,t are mappings of F into S.

Let C=(S,F,o,t)C = (S,F, o, t) be a quiver. The elements of S are called the vertices of C. The elements of F are called the arrows of C. Let ff be an arrow of C; the vertex o(f)o(f) is called the origin or the source of ff, the vertex t(f)t(f) is called the end, or the target, of ff; one also says that ff joins the vertex o(f)o(f) to the vertex t(f)t(f). The mappings oo and tt are called respectively the source and target mappings, or origin and end, of the quiver C.

© N. Bourbaki and Springer-Verlag Berlin Heidelberg 2 016

N. Bourbaki, Algebraic Topology, DOI 10.1007/978-3-662-49361-8_2 151

When C is a quiver, Som(C), Fl(C), oCo_C and tCt_C will denote respectively the set of vertices, the set of arrows, the source mapping and the target mapping of C. Let aa and bb be vertices of C; the set of arrows of C joining aa to bb is denoted by Fla,b(C)_{a,b}(C), or again Ca,bC_{a,b}.

To make the arguments easier to understand, one may represent a quiver by a diagram composed of points and arrowed lines corresponding respectively to the vertices and the arrows. The diagram:

f1g f_1g
∙∙∙k\bullet \bullet \bullet k abcabc f2hf_2h

represents a quiver whose vertices are a,b,ca,b,c, whose arrows are f1,f2,g,h,kf_1, f_2, g, h, k, and whose source and target mappings are given by

o(f1)=o(f2)=ao(g)=bo(h)=co(k)=c o(f_1) =o(f_2) =ao(g) =bo(h) =co(k) =c
t(f1)=t(f2)=bt(g)=ct(h)=bt(k)=c t(f_1) =t(f_2) =bt(g) =ct(h) =bt(k) =c

2. Subquivers

Let C and C′C' be quivers. One says that C′C' is a subquiver of C if one has Som(C′')⊂\subset Som(C), Fl(C′')⊂\subset Fl(C) and if the mappings oC′o_{C'} and tC′t_{C'} coincide on Fl(C′') with the mappings oCo_C and tCt_C. If, moreover, every arrow of C whose source and target belong to Som(C′') is an arrow of C′C', one says that C′C' is a full subquiver of C.

Let C be a quiver and let (Ci)i∈I(C_i)_{i\in I} be a family of subquivers of C. The intersection of the family (Ci)i∈I(C_i)_{i\in I} is defined to be, and is denoted by ⋂i∈ICi\bigcap_{i\in I}C_i, the subquiver of C whose set of vertices is ⋂i∈I\bigcap_{i\in I} Som(Ci_i) and whose set of arrows is ⋂i∈I\bigcap_{i\in I} Fl(Ci_i).

3. Morphisms of quivers

Let C and C′C' be quivers. A quiver morphism of C into C′C' is a pair (u,v)(u, v), where u:u: Som(C) →\rightarrow Som(C′') and v:v: Fl(C) →\rightarrow Fl(C′') are mappings such that oC′∘v=u∘oCo_{C'}\circ v=u\circ o_C and tC′∘v=u∘tCt_{C'}\circ v=u\circ t_C.

Let φ=(u,v)\varphi = (u, v) be a quiver morphism of C into C′C'. The mappings uu and vv are denoted by Som(φ\varphi ) and Fl(φ\varphi ). If aa is a vertex of C, we shall denote by abuse φ(a)\varphi (a) the image of a vertex aa of C, and we shall say that φ(a)\varphi (a) is the image of aa under φ\varphi. Analogously, if ff is an arrow of C, we shall denote by abuse φ(f)\varphi (f) the arrow v(f)v(f) of C′C' and we shall say that it is the image of ff under φ\varphi.

Let C, C′,C′′C',C'' be three quivers, let φ=(u,v)\varphi = (u, v) be a morphism of C into C′C' and let φ′=(u′,v′)\varphi '= (u', v') be a morphism of C′C' into C′′C''. Then (u′∘u,v′∘v)(u'\circ u, v'\circ v) is a morphism of C into C′′C'' which is denoted by φ′∘φ\varphi '\circ \varphi and is called the composite of φ′\varphi ' and φ\varphi.

IdC_C denotes the morphism (IdSom(C)_{Som(C)}, IdFl(C)_{Fl(C)}) of C into itself.

Let φ\varphi be a quiver morphism of C into C′C'. In order that φ\varphi be an isomorphism, it is necessary and sufficient that the mappings Som(φ\varphi ) and Fl(φ\varphi ) be bijective (cf. E, IV, p. 6).

Thus, if one calls a quiver structure on sets S and F the datum of two mappings o,t:F→So, t: F\rightarrow S, one may take quiver morphisms as morphisms of this structure (E, IV, p. 11).

Let φ\varphi be a quiver morphism of C into C′C'. There exists a unique subquiver of C′C' whose sets of vertices and arrows are the images of Som(φ\varphi ) and Fl(φ\varphi ) respectively. It is denoted by φ(C)\varphi (C) and is called the image of C under φ\varphi. One says that φ\varphi is injective if the mappings Som(φ\varphi ) and Fl(φ\varphi ) are injective; in this case, φ\varphi induces an isomorphism of C onto its image.

Let φ\varphi and ψ\psi be morphisms of quivers from C to C′C'. There exists a unique subquiver of C whose vertices and arrows are respectively the vertices and arrows of C having the same image under φ\varphi and ψ\psi. It is called the equaliser of φ\varphi and ψ\psi.

4. Products of quivers

Let (Ci)i∈I(C_i)_{i\in I} be a family of quivers. Put S = ∏i∈I\prod_{i\in I} Som(Ci_i), F=∏i∈IF =\prod_{i\in I} Fl(Ci_i), and let oo and tt be the mappings ∏i∈IoCi\prod_{i\in I}o_{C_i} and ∏i∈ItCi\prod_{i\in I}t_{C_i} respectively. The quadruple C=(S,F,o,t)C = (S,F, o, t) is a quiver, called the product quiver of the family (Ci)i∈I(C_i)_{i\in I}.

There exists a unique morphism of quivers pri:C→Ci_i: C\rightarrow C_i such that the mappings Som(pri_i) and Fl(pri_i) are respectively the projections of index ii from Som(C) to Som(Ci_i) and from Fl(C) to Fl(Ci_i).

We have the following universal property: Let C′C' be a quiver; for every family (φi)i∈I(\varphi_i)_{i\in I}, where, for every i∈I,φi:C′→Cii\in I,\varphi_i: C'\rightarrow C_i is a morphism of quivers, there exists a unique morphism of quivers φ:C′→C\varphi : C'\rightarrow C such that φi=\varphi_i= pri∘φ_i\circ \varphi for every i∈Ii\in I.

5. Paths and loops in a quiver

Definition 2 01ST

Let C be a quiver. A path in C is a sequence c=(a0,f1,a1,...,an−1,fn,an)c= (a_0, f_1, a_1, . . . , a_{n-1}, f_n, a_n), where nn is an integer ⩾0,a0,...,an\geqslant 0,a_0, . . . , a_n are vertices of C, and where, for 1⩽i⩽n,fi1\leqslant i\leqslant n,f_i is an arrow of C joining ai−1a_{i-1} to aia_i. One says that the path cc is of length nn.

Let c=(a0,f1,a1,...,an−1,fn,an)c= (a_0, f_1, a_1, . . . , a_{n-1}, f_n, a_n) be a path of length nn in C. The vertex a0a_0 is called the origin, or the source, of cc; the vertex ana_n is called the term, or the end, of cc; one also says that cc joins the vertex a0a_0 to the vertex ana_n. The vertex aia_i, for 0⩽i⩽n0\leqslant i\leqslant n, is called the vertex of index ii; the arrow fif_i, for 1⩽i⩽n1\leqslant i\leqslant n, is called the ii-th arrow, or the arrow of index ii. A path of length ⩾1\geqslant 1 is determined by the sequence of its arrows.

A path whose origin is equal to the term is called a loop. A path of length 0 is a loop, called constant.

One says that paths

c=(a0,f1,a1,...,an−1,fn,an)c= (a_0, f_1, a_1, . . . , a_{n-1}, f_n, a_n) and c′=(a0′,f1′,a1′,...,am−1′,fm′,am′)c'= (a'_0, f'_1, a'_1, . . . , a'_{m-1}, f'_m, a'_m)

in C are juxtaposed if the term ana_n of cc is the origin a0′a'_0 of c′c'. In this case, the sequence (a0,f1,a1,...,an−1,fn,an,f1′,a1′,...,fm′,am′)(a_0, f_1, a_1, . . . , a_{n-1}, f_n, a_n, f'_1, a'_1, . . . , f'_m, a'_m) is a path in C which is denoted by c∗c′c*c' and is called the juxtaposed path of cc and c′c'. It joins the origin of cc to the term of c′c'; its length is the sum of the lengths of the paths cc and c′c'.

6. Connected components of a quiver

Let C=(S,F,o,t)C = (S,F, o, t) be a quiver. Consider the finest equivalence relation RCR_C in S such that the relation RC{o(f),t(f)}R_C\{o(f), t(f)\} is satisfied for every arrow ff of C. Two vertices a,ba,b of C are equivalent for this relation if and only if there exist an integer n⩾0n\geqslant 0, vertices a0,...,ana_0, . . . , a_n of C and arrows f1,...,fnf_1, . . . , f_n of C such that a0=a,an=ba_0=a,a_n=b and such that, for 1⩽i⩽n1\leqslant i\leqslant n, the arrow fif_i joins either ai−1a_{i-1} to aia_i, or aia_i to ai−1a_{i-1}.

The equivalence classes of this equivalence relation are called the connected components of C. We denote by π0(C)\pi_0(C) the set of connected components of C. Finally, we say that C is connected if it has at most one connected component.

Let φ:C→C′\varphi : C\rightarrow C' be a morphism of quivers. The map Som(φ\varphi ) defines by passing to the quotients a map of π0(C)\pi_0(C) into π0(C′)\pi_0(C') which we shall denote by π0(φ)\pi_0(\varphi ).

Exercises

See the exercises for § 1.

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