• Aucun résultat trouvé

3.9 Substituting the Axiom of Inverse

3.9.1 Specifying Domains of Existence

Thus, we considerK-analytic transformation equationsx0i= fi(x;a)defined on a more general domain than a product∆ρn×rσof two small polydiscs centered at the origin. Here is how Lie and Engel specify their domains of existence on p. 14 of [1], and these domains might be global.

§ 2.

In the transformation equations:

(1) x0i=fi(x1, . . . ,xn;a1, . . . ,ar) (i=1···n),

let now all the parametersa1, . . . ,arbe essential.

Since the fi are analytic functions of their arguments, in the domain [GE

-BIETE] of all systems of valuesx1, . . . ,xnand in the domain of all systems of valuesa1, . . . ,ar, we can choose a region [BEREICH](x)and, respectively, a region(a)such that the following holds:

Firstly. The fi(x,a)are single-valued [EINDEUTIG] functions of then+r variablesx1, . . . ,xn,a1, . . . ,arin the complete extension [AUSDEHNUNG] of the two regions(x)and(a).

Secondly. Thefi(x,a)behave regularly in the neighbourhood of every sys-tem of valuesx01, . . . ,x0n,a01, . . . ,a0r, hence are expandable in ordinary power se-ries with respect tox1−x01, . . . ,xn−x0n, as soon asx01, . . . ,x0nlies arbitrarily in the domain(x), anda01, . . . ,a0r lies arbitrarily in the domain(a).

Thirdly. The functional determinant:

±xf11···xfnn

vanishes for no combination of systems of values of xi and ofak in the two domains(x)and(a), respectively.

Fourthly. If one gives to the parametersakin the equationsx0i=fi(x,a)any of the valuesa0kin domain(a), then the equations:

x0i=fi(x1, . . . ,xn;a01, . . . ,a0r) (i=1···n)

always produce two different systems of valuesx01, . . . ,x0nfor two different sys-tems of valuesx1, . . . ,xnof the domain(x).

We assume that the two regions(x)and(a)are chosen in such a way that these four conditions are satisfied. If we give to the variablesxiin the equations x0i=fi(x,a)all possible values in(x)and to the parametersakall possible values in(a), then in their domain, thex0irun throughout a certain region, which we can denote symbolically by the equationx0= f (x)(a)

. This new domain has the following properties:

Firstly. Ifa01, . . . ,a0ris an arbitrary system of values of(a)andx010, . . . ,x0n0an arbitrary system of values of the subregionx0=f (x)(a0)

, then in the neigh-bourhood of the system of valuesx0i0,a0k, thex1, . . . ,xncan be expanded in ordi-nary power series with respect tox01−x010, . . . ,x0n−x0n0,a1−a01, . . . ,an−a0n.

Secondly. If one gives to theakfixed valuesa0k in domain(a), then in the equations:

x0i=fi(x1, . . . ,xn;a01, . . . ,a0r) (i=1···n),

the quantitiesx1, . . . ,xnwill be single-valued functions ofx01, . . . ,x0n, which be-have regularly in the complete extension of the regionx0= f (x)(a0)

.

Referring also to the excerpt p. 26, we may rephrase these basic assumptions as follows. The fi(x;a)are defined for(x,a)belonging to the product:

X ×A Kn×Kr

of two domainsX KnandA Kr. These functions areK-analytic in both vari-ables, hence expandable in Taylor series at every point(x0,a0). Furthermore, for every fixeda0, the mapx7→ f(x;a0)is assumed to constitute aK-analytic diffeo-morphism ofX × {a0}onto its image. Of course, the inverse map is also locally expandable in power series, by virtue of theK-analytic inverse function theorem.

To insure that the composition of two transformations exists, one requires that there exist nonempty subdomains:

X1⊂X and A1⊂A, with the property that for every fixeda1∈A1:

f X1× {a1}

⊂X,

so that for every such ana1∈A1and for every fixedb∈A, the composed map:

x7−→f f(x;a1);b

is well defined for allx∈X1and moreover, is aK-analytic diffeomorphism onto its image. In fact, it is evenK-analytic with respect to all the variables(x,a1,b)in X1×A1×A. Lie’s fundamental and uniquegroup composition axiommay then be expressed as follows.

(A1) There exists aKr-valuedK-analytic mapϕ=ϕ(a,b)defined inA1×A1with ϕ A1×A1

⊂A such that:

f f(x;a);b

f x;ϕ(a,b) for all x∈X1,a∈A1, b∈A1.

3.9 Substituting the Axiom of Inverse for a Differential Equations Assumption 53 Here are two further specific unmentioned assumptions that Lie presupposes, still with the goal of admitting neither the identity element, nor the existence of inverses.

(A2) There is aKr-valuedK-analytic map(a,c)7−→b=b(a,c)defined forarunning in a certain (nonempty) subdomain A2⊂A1 and for c running in a certain (nonempty) subdomainC2⊂A1withb A2×C2

⊂A1 which solvesb in terms of(a,c)in the equationsckk(a,b), namely which satisfies identically:

c≡ϕ a,b(a,c) for all a∈A2,c∈C2.

Inversely,ϕ(a,b)solvescin terms of(a,b)in the equationsbk=bk(a,c), namely more precisely: there exists a certain (nonempty) subdomainA3⊂A2(⊂A1) and a certain (nonempty) subdomainB3⊂A1withϕ A3×B3

⊂C2such that one has identically:

b≡b a,ϕ(a,b) for all a∈A3, b∈B3.

Example.In Engel’s counterexample of the groupx0(λ)xwith a Riemann uniformizing mapω :Λ as on p. 21 having inverseχ:Λ ∆, these three requirements are satisfied, and in addition, we claim that one may even takeX = KandA1=A =Λ, with no shrinking, for composition happens to hold in fact without restriction in this case. Indeed, starting from the general composition:

x00(λ2)x0(λ2(λ1)x,

that is to say, fromx00(λ2(λ1)x, in order to represent it under the specific form x00(λ3)x, it is necessary and sufficient to solveχ(λ3) =χ(λ1(λ2), hence we may take forϕ:

λ3=ω χ(λ1(λ2)

=:ϕ(λ1,λ2),

without shrinking the domains, for the two inequalities|χ(λ1)|<1 and|χ(λ2)|<1 readily imply that|χ(λ1(λ2)|<1 too so thatω χ(λ1(λ2)

is defined. On the other hand, for solvingλ2in terms of(λ1,λ3)in the above equation, we are naturally led to define:

b(λ1,λ3):=ω χ(λ3) χ(λ1)

,

and then b=b(λ1,λ3) is defined under the specific restriction that |χ(λ3)|<

|χ(λ1)|.

However, the axiom (A2)happens to be still incomplete for later use, and one should add the following axiom in order to be able to also solveainc(a,b).

(A3)There is aKm-valuedK-analytic map(b,c)7−→a=a(b,c)defined inB4×C4 withB4⊂A1andC4⊂A1, and witha B4×C4

⊂A1, such that one has identically:

c≡ϕ a(b,c),b for all b∈B4, c∈C4.

Inversely,ϕ(a,b)solvescin the equationsak=ak(b,c), namely more precisely:

there existB5⊂B4andA5⊂A1withϕ A5×B5

⊂C4such that one has

identically:

a≡a b,ϕ(a,b) for all a∈A5, b∈B5, and furthermore in addition, witha B5×C4

⊂A2and withb A5×C4

B4such that one also has identically:

b≡b a(b,c),c

for all b∈B5, c∈C4 a≡a b(a,c),a

for all a∈A5,c∈C4.

The introduction of the numerous (nonempty) domainsA2,C2,A3,B3,B4, C4,A5,B5which appears slightly unnatural and seems to depend upon the order in which the solving mapsaandbare considered can be avoided by requiring from the beginning only that there exist two subdomainsA3⊂A2⊂A1such that one has uniformly:

c≡ϕ a,b(a,c) for all a∈A3, c∈A3 c≡ϕ a(b,c),b for all b∈A3,c∈A3 b≡b a(b,c),c

for all b∈A3,c∈A3 b≡b a,ϕ(a,b) for all a∈A3,b∈A3 a≡a b(a,c),c

for all a∈A3,c∈A3 a≡a b,ϕ(a,b) for all a∈A3, b∈A3.

We will adopt these axioms in the next subsection. Importantly, we would like to point out that, although b(a,c) seems to represent the group product a−1·c=m i(a),c

, the assumption (A2) neither reintroduces inverses, nor the identity element, it just means that one may solvebby means of the implicit function theorem in the parameter composition equationsckk(a1, . . . ,ar,b1, . . . ,br).

3.9.2 Group Composition Axiom And Fundamental Differential

Documents relatifs