Bourbaki

Proposition 6

Tag 01UH, permanent. Cite it as [Bourbaki, Tag 01UH].

Topologie algébrique, chapter II, § 4. Homotopies

There exists a unique group homomorphism

Λ:(∗Ga(i))∗\Lambda :(*G_{a(i)})* F(Orb(H)) →\rightarrow Coh(φ,ψ\varphi , \psi )a(i0)_{a(i_0)}

i∈i\inOrb(G)

such that

(7) Λ(f)=δiαi(f)δi−1\Lambda (f) =\delta_i\alpha_i(f)\delta_i^{-1} for i∈i\in Orb(G) and f∈Ga(i)f\in G_{a(i)},

(8) Λ(j)=δφ0(j)τ0(j)δψ0(j)−1\Lambda (j) =\delta_{\varphi_0(j)}\tau_0(j)\delta^{-1}_{\psi_0(j)} for j∈j\in Orb(H).

The homomorphism Λ is surjective; its kernel is the smallest normal subgroup of (∗iGa(i))∗(*_iG_{a(i)})* F(Orb(H)) containing the elements jj of Fl(T) and the elements φj(f)jψj(f)−1j−1\varphi_j(f)j\psi_j(f)^{-1}j^{-1}, for j∈j\in Orb(H) and f∈Hb(j)f\in H_{b(j)}.

The existence and uniqueness of the homomorphism Λ result from the universal property of free products and free groups (A, I, p. 85, prop. 8).

Let A denote the set of the a(i)a(i) for i∈i\in Orb(G) and GAG_A the full subgroupoid of G whose set of vertices is A. For every x∈x\in Som(G), let x‾\overline{x} denote the orbit of xx in G and choose an arrow dxd_x of G joining xx to a(x‾)a(\overline{x}). The pair vv formed by the mapping x↦a(x‾)x\mapsto a(\overline{x}) of Som(G) into A and the mapping which to f∈f\in Flx,y(G)_{x,y}(G) associates the element dx−1fdyd^{-1}_xf d_y of Fla(x‾),a(y‾)(GA)_{a(\overline{x}),a(\overline{y})}(G_A) is a groupoid morphism. It follows from Proposition 1 of II, p. 182 that vv is a homotopism. Let φ′=v∘φ\varphi '=v\circ \varphi and ψ′=v∘ψ\psi '=v\circ \psi, and then let ww be the canonical morphism of Coh(φ,ψ\varphi , \psi ) into Coh(φ′,ψ′\varphi ', \psi ') ; it is a homotopism (II, p. 187, Theorem 1).

The orbits of GAG_A are the sets {a}\{a\}, for a∈Aa\in A, and the injection GA→GG_A\rightarrow G induces a bijection of Orb(GA_A) onto Orb(G), by means of which we shall identify these two sets. One defines a basic equipment (a′,b′,β1′,β2′,T′,i0)(a', b', \beta '_1, \beta '_2,T', i_0) of the pair (φ′,ψ′)(\varphi ', \psi ') by setting a′(i)=a(i)a'(i) =a(i) for i∈i\in Orb(G),b′(j)=b(j),β1′(j)=v(c1(j)),β2′(j)=v(c2(j)),b'(j) =b(j),\beta '_1(j) =v(c_1(j)),\beta '_2(j) =v(c_2(j)) for j∈j\in Orb(H) and T′=TT'= T. The group homomorphisms φj′\varphi '_j and ψj′\psi '_j (for j∈j\in Orb(H)), the quiver morphism τ0′\tau '_0, the arrows δi′\delta '_i (for i∈i\in Orb(G)), and therefore the group homomorphism Λ′\Lambda ', deduced from this basic equipment, are the composites with ww of the corresponding homomorphisms φj,ψj\varphi_j,\psi_j, of the quiver morphism τ0\tau_0, of the corresponding arrows δi\delta_i and of the homomorphism Λ.

Let B be the set of the b(j)b(j) for j∈j\in Orb(H), let HBH_B be the full subgroupoid of H with set of vertices B; let us denote by u:HB→Hu: H_B\rightarrow H the canonical injection; put φ′′=φ′∘u\varphi ''=\varphi '\circ u and ψ′′=ψ′∘u\psi ''=\psi '\circ u. The morphism uu induces a bijection B →\rightarrow Orb(H) by which we shall identify these two sets. From theorem 1 of II, p. 187 one further deduces a canonical homotopism w′:w': Coh(φ′′,ψ′′\varphi '', \psi '')→\rightarrow Coh(φ′,ψ′\varphi ', \psi '). Moreover, the pair (φ′′,ψ′′)(\varphi '', \psi '') is endowed with a base equipment (a′′,b′′,β1′′,β2′′,T′′,i0)(a'', b'', \beta ''_1, \beta ''_2,T'', i_0), so that Λ′,φj′,ψj′,τ0′,δi′(i,i′∈\Lambda ', \varphi '_j, \psi '_j, \tau '_0, \delta '_i(i, i'\in Orb(G), j∈j\in Orb(H)) are the composites with w′w' of Λ′′,φj′′,ψj′′,τ0′′,δi′′\Lambda '', \varphi ''_j, \psi ''_j, \tau ''_0, \delta ''_i.

The various morphisms of groupoids introduced are summarized by the following diagram:

HBψ′′φ′′GAα′′H_{B\psi''}^{\varphi''}G_A^{\alpha''} Coh(φ′′,ψ′′\varphi '', \psi '')

w′w'

(9) H φ′GAα′^{\varphi'}G_A^{\alpha'} Coh(φ′′,ψ′′\varphi '', \psi '')

ψ′\psi '
vw vw

H ψφ^{\varphi}_{\psi} G α^{\alpha} Coh(φ,ψ\varphi , \psi ) .

To prove the proposition, we may therefore suppose that A = Som(G) and B = Som(H), in other words that the canonical mappings Som(G) →\rightarrow Orb(G) and Som(H) →\rightarrow Orb(H) are bijective, hypotheses under which we shall place ourselves in the rest of the proof.

The quiver G1G_1 then has for set of vertices A and for arrows the sum set of the sets Ga,a∈AG_a,a\in A, and of the set B. The arrows of GaG_a are loops at aa; if b∈Bb\in B, the arrow bb joins φ(b)\varphi (b) to ψ(b)\psi (b), the arrows c1(b)c_1(b) and c2(b)c_2(b) are loops respectively at φ(b)\varphi (b) and ψ(b)\psi (b). The quiver T will be identified with an oriented tree of G1G_1; it is a maximal oriented tree since the set of its vertices is equal to the set of vertices of G1(G_1(II, p. 157, prop. 1). We put a0=a(i0)a_0=a(i_0). The set of arrows of the framework Γ0\Gamma_0 of the pair (φ,ψ)(\varphi , \psi ) being identified with B, the morphism of quivers τ1:Γ0→\tau_1: \Gamma_0\rightarrow Grp(G1_1) associates to the arrow bb the class of paths c1(b)−1bc2(b)c_1(b)^{-1}bc_2(b) in the graph G~1\widetilde{G}_1.

Recall that θ1\theta_1 denotes the canonical morphism of quivers of G1G_1 into Grp(G1_1). Let

λ:a∗∈AF(Ga)∗F(B)→\lambda :_a*_{\in A}F(G_a)*F(B)\rightarrow Grp(G1_1)a0_{a_0}

be the unique group homomorphism such that one has

λ(f)=τ1(da)θ1(f)τ1(da)−1\lambda (f) =\tau_1(d_a)\theta_1(f)\tau_1(d_a)^{-1} if a∈Aa\in A and f∈Gaf\in G_a;

λ(b)=τ1(dφ(b))τ1(b)τ1(dψ(b))−1\lambda (b) =\tau_1(d_{\varphi(b)})\tau_1(b)\tau_1(d_{\psi(b)})^{-1} if b∈Bb\in B.

We thus have λ=λ′∘ε\lambda =\lambda '\circ \varepsilon, where λ′\lambda ' denotes the canonical group homomorphism from a∗∈AF(Ga)∗F(B)_a*_{\in A}F(G_a)*F(B) into Grp(G1_1)a0_{a_0} defined by the maximal oriented tree T (II, p. 179, prop. 9), and where ε\varepsilon is the unique automorphism of the group a∗∈AF(Ga)∗F(B)_a*_{\in A}F(G_a)*F(B) such that ε(f)=f\varepsilon (f) =f for a∈Aa\in A and f∈Gaf\in G_a, and ε(b)=c1(b)−1h1(b)c2(b)\varepsilon (b) =c_1(b)^{-1}h_1(b)c_2(b) for b∈Bb\in B. According to remark 2 of II, p. 179, the homomorphism λ\lambda is surjective and its kernel is the smallest normal subgroup of a∗∈AF(Ga)∗F(B)_a*_{\in A}F(G_a)*F(B) containing the arrows of T.

Let us denote by π:\pi : Grp(G1_1)→\rightarrow Coh(φ,ψ\varphi , \psi ) the canonical groupoid morphism. According to II, p. 177, cor. 1 of prop. 8, the group homomorphism πa0\pi_{a_0} from Grp(G1_1)a0_{a_0} into Coh(φ,ψ\varphi , \psi )a0_{a_0} is surjective, and its kernel is the smallest normal subgroup of Grp(G1_1)a0_{a_0} containing the loops Int(τ1(da)\tau_1(d_a))(α1(f)α1(g)α1(fg)−1)(\alpha_1(f)\alpha_1(g)\alpha_1(f g)^{-1}), for a∈Aa\in A and f,g∈Gaf,g\in G_a, and the loops Int(τ1(dφ(b))\tau_1(d_{\varphi(b)}))(φ(f)bψ(f)−1b−1)(\varphi (f)b\psi (f)^{-1}b^{-1}), for b∈Bb\in B and f∈Hbf\in H_b.

If p:F(⋃Ga∪B)→(a∗∈AGa)∗F(B)p: F(\bigcup G_a\cup B)\rightarrow (_a*_{\in A}G_a)*F(B) denotes the canonical surjective homomorphism, we thus have Λ∘p=πa0∘λ\Lambda \circ p=\pi_{a_0}\circ \lambda. This formula implies that the homomorphism Λ is surjective; it remains to determine its kernel.

For a∈Aa\in A and f∈Gaf\in G_a, let us denote by [f][f] the image of f∈F(Ga)f\in F(G_a) in the group a∗∈AF(Ga)∗F(B)_a*_{\in A}F(G_a)*F(B). For a∈A,f,g∈Gaa\in A,f, g\in G_a, we then have

Int(τ1(da)\tau_1(d_a))(α1(f)α1(g)α1(fg)−1)=λ([f][g][fg]−1)(\alpha_1(f)\alpha_1(g)\alpha_1(f g)^{-1}) =\lambda ([f][g][f g]^{-1}).

Analogously, for b∈Bb\in B and f∈Hbf\in H_b, the definition of the homomorphisms φb\varphi_b and ψb\psi_b (formula (6) of II, p. 193) implies that one has

Int(τ1(dφ(b))\tau_1(d_{\varphi(b)}))(φ(f)h1(b)ψ(f)−1h1(b)−1)(\varphi (f)h_1(b)\psi (f)^{-1}h_1(b)^{-1})

=τ1(dφ(b))φ(f)(c1(b)τ1(b)c2(b)−1)ψ(f)−1 =\tau_1(d_{\varphi(b)})\varphi (f)(c_1(b)\tau_1(b)c_2(b)^{-1})\psi (f)^{-1}
(c2(b)τ1(b)−1c1(b)−1)τ1(dφ(b))−1 (c_2(b)\tau_1(b)^{-1}c_1(b)^{-1})\tau_1(d_{\varphi(b)})^{-1}
=τ1(dφ(b))c1(b)φb(f)τ1(b)ψb(f)−1τ1(b)−1c1(b)−1τ1(dφ(b))−1 =\tau_1(d_{\varphi(b)})c_1(b)\varphi_b(f)\tau_1(b)\psi_b(f)^{-1}\tau_1(b)^{-1}c_1(b)^{-1}\tau_1(d_{\varphi(b)})^{-1}
=λ(c1(b))λ(φb(f)[b]ψb(f)−1[b]−1)λ(c1(b))−1 =\lambda (c_1(b))\lambda (\varphi_b(f)[b]\psi_b(f)^{-1}[b]^{-1})\lambda (c_1(b))^{-1}

Consequently, the kernel of the homomorphism πa0∘λ\pi_{a_0}\circ \lambda is the smallest normal subgroup of a∗∈AF(Ga)∗F(B)_a*_{\in A}F(G_a)*F(B) which contains the elements [f][g][fg]−1[f][g][f g]^{-1} for a∈Aa\in A and f,g∈Gaf,g\in G_a, the elements φb(f)[b]ψb(f)−1[b]−1\varphi_b(f)[b]\psi_b(f)^{-1}[b]^{-1} for b∈Bb\in B and f∈Hbf\in H_b, and the elements [b][b], for b∈b\in Fl(T).

Finally, the kernel of the homomorphism Λ is the smallest normal subgroup of (a∗∈AGa)∗F(B)(_a*_{\in A}G_a)*F(B) which contains the images under pp of the preceding elements, in other words, the elements [b][b], for b∈b\in Fl(T), and the elements φb(f)[b]ψb(f)[b]−1\varphi_b(f)[b]\psi_b(f)[b]^{-1}, for b∈Bb\in B and f∈Hbf\in H_b. The proposition is thus proved.

Bourbaki sets the proof directly after the statement, and the corpus does not separate them, so what is above is the statement and its proof together.

Cited by

Nothing.

Cites

Nothing.

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