Definition 2.1 .

Let H 𝐻 H italic_H be a Lie group and G 𝐺 G italic_G a Lie subgroup of H 𝐻 H italic_H . Denote by π”₯ π”₯ \mathfrak{h} fraktur_h the Lie algebra of H 𝐻 H italic_H and by 𝔀 𝔀 \mathfrak{g} fraktur_g the Lie algebra of G 𝐺 G italic_G . We shall say that G 𝐺 G italic_G is a reductive Lie subgroup of H 𝐻 H italic_H if there exists a direct sum decomposition

π”₯ = 𝔀 βŠ• π”ͺ , π”₯ direct-sum 𝔀 π”ͺ \mathfrak{h}=\mathfrak{g}\oplus\mathfrak{m}, fraktur_h = fraktur_g βŠ• fraktur_m ,

where π”ͺ π”ͺ \mathfrak{m} fraktur_m is an Ad G subscript Ad 𝐺 \mathrm{Ad}_{G} roman_Ad start_POSTSUBSCRIPT italic_G end_POSTSUBSCRIPT -invariant vector subspace of π”₯ π”₯ \mathfrak{h} fraktur_h , i.e. Ad a ⁒ ( π”ͺ ) βŠ‚ π”ͺ subscript Ad π‘Ž π”ͺ π”ͺ \mathrm{Ad}_{a}(\mathfrak{m})\subset\mathfrak{m} roman_Ad start_POSTSUBSCRIPT italic_a end_POSTSUBSCRIPT ( fraktur_m ) βŠ‚ fraktur_m for all a ∈ G π‘Ž 𝐺 a\in G italic_a ∈ italic_G (which means that the Ad G subscript Ad 𝐺 \mathrm{Ad}_{G} roman_Ad start_POSTSUBSCRIPT italic_G end_POSTSUBSCRIPT representation of G 𝐺 G italic_G in π”₯ π”₯ \mathfrak{h} fraktur_h is reducible into a direct sum decomposition of two Ad G subscript Ad 𝐺 \mathrm{Ad}_{G} roman_Ad start_POSTSUBSCRIPT italic_G end_POSTSUBSCRIPT -invariant vector spaces: cf. [ 18 ] , p. 83 83 83 83 ).

Definition 2.7 .

Let P ⁒ ( M , G ) 𝑃 𝑀 𝐺 P(M,G) italic_P ( italic_M , italic_G ) be a principal bundle and ρ : Ξ“ β†’ G : 𝜌 β†’ Ξ“ 𝐺 \rho\colon\Gamma\to G italic_ρ : roman_Ξ“ β†’ italic_G a central homomorphism of a Lie group Ξ“ Ξ“ \Gamma roman_Ξ“ onto G 𝐺 G italic_G , i.e.Β such that its kernel is discrete and contained in the centre of Ξ“ Ξ“ \Gamma roman_Ξ“ [ 14 ] (see also [ 15 ] ). A Ξ“ normal-Ξ“ \Gamma roman_Ξ“ -structure on P ⁒ ( M , G ) 𝑃 𝑀 𝐺 P(M,G) italic_P ( italic_M , italic_G ) is a principal bundle map ΞΆ : P ~ β†’ P : 𝜁 β†’ ~ 𝑃 𝑃 \zeta\colon\tilde{P}\to P italic_ΞΆ : ~ start_ARG italic_P end_ARG β†’ italic_P which is equivariant under the right actions of the structure groups, i.e.

ΞΆ ⁒ ( u ~ β‹… Ξ± ) = ΞΆ ⁒ ( u ~ ) β‹… ρ ⁒ ( Ξ± ) 𝜁 β‹… ~ 𝑒 𝛼 β‹… 𝜁 ~ 𝑒 𝜌 𝛼 \zeta(\tilde{u}\cdot\alpha)=\zeta(\tilde{u})\cdot\rho(\alpha) italic_ΞΆ ( ~ start_ARG italic_u end_ARG β‹… italic_Ξ± ) = italic_ΞΆ ( ~ start_ARG italic_u end_ARG ) β‹… italic_ρ ( italic_Ξ± )

for all u ~ ∈ P ~ ~ 𝑒 ~ 𝑃 \tilde{u}\in\tilde{P} ~ start_ARG italic_u end_ARG ∈ ~ start_ARG italic_P end_ARG and Ξ± ∈ Ξ“ 𝛼 Ξ“ \alpha\in\Gamma italic_Ξ± ∈ roman_Ξ“ .


Definition 1.1 (Zettl [ 38 ] ) .

A C*-ternary ring is a Banach space A 𝐴 A italic_A with ternary product [ x , y , z ] : A Γ— A Γ— A β†’ A normal-: π‘₯ 𝑦 𝑧 normal-β†’ 𝐴 𝐴 𝐴 𝐴 [x,y,z]:A\times A\times A\rightarrow A [ italic_x , italic_y , italic_z ] : italic_A Γ— italic_A Γ— italic_A β†’ italic_A which is linear in the outer variables, conjugate linear in the middle variable, is associative :

[ a ⁒ b ⁒ [ c ⁒ d ⁒ e ] ] = [ a ⁒ [ d ⁒ c ⁒ b ] ⁒ e ] = [ a ⁒ b ⁒ [ c ⁒ d ⁒ e ] ] , delimited-[] π‘Ž 𝑏 delimited-[] 𝑐 𝑑 𝑒 delimited-[] π‘Ž delimited-[] 𝑑 𝑐 𝑏 𝑒 delimited-[] π‘Ž 𝑏 delimited-[] 𝑐 𝑑 𝑒 [ab[cde]]=[a[dcb]e]=[ab[cde]], [ italic_a italic_b [ italic_c italic_d italic_e ] ] = [ italic_a [ italic_d italic_c italic_b ] italic_e ] = [ italic_a italic_b [ italic_c italic_d italic_e ] ] ,

and satisfies βˆ₯ [ x ⁒ y ⁒ z ] βˆ₯ ≀ βˆ₯ x βˆ₯ ⁒ βˆ₯ y βˆ₯ ⁒ βˆ₯ z βˆ₯ norm delimited-[] π‘₯ 𝑦 𝑧 norm π‘₯ norm 𝑦 norm 𝑧 \|[xyz]\|\leq\|x\|\|y\|\|z\| βˆ₯ [ italic_x italic_y italic_z ] βˆ₯ ≀ βˆ₯ italic_x βˆ₯ βˆ₯ italic_y βˆ₯ βˆ₯ italic_z βˆ₯ and βˆ₯ [ x ⁒ x ⁒ x ] βˆ₯ = βˆ₯ x βˆ₯ 3 norm delimited-[] π‘₯ π‘₯ π‘₯ superscript norm π‘₯ 3 \|[xxx]\|=\|x\|^{3} βˆ₯ [ italic_x italic_x italic_x ] βˆ₯ = βˆ₯ italic_x βˆ₯ start_POSTSUPERSCRIPT 3 end_POSTSUPERSCRIPT .


Definition 2.1 .

If k , m π‘˜ π‘š k,m italic_k , italic_m are positive integers, we define their unitary product as

k βŠ• m = { k ⁒ m gcd ⁑ ( k , m ) = 1 0 otherwise direct-sum π‘˜ π‘š cases π‘˜ π‘š π‘˜ π‘š 1 0 otherwise k\oplus m=\begin{cases}km&\gcd(k,m)=1\\ 0&\text{ otherwise}\end{cases} italic_k βŠ• italic_m = { start_ROW start_CELL italic_k italic_m end_CELL start_CELL roman_gcd ( italic_k , italic_m ) = 1 end_CELL end_ROW start_ROW start_CELL 0 end_CELL start_CELL otherwise end_CELL end_ROW (4)

If k βŠ• m = p direct-sum π‘˜ π‘š 𝑝 k\oplus m=p italic_k βŠ• italic_m = italic_p , then we write k | | p fragments k fragments | fragments | p {k\left\lvert\lvert{p}\right.} italic_k | | italic_p and say that k π‘˜ k italic_k is a unitary divisor of p 𝑝 p italic_p .


Definition 2.3

A smooth curve c : I β†’ N : 𝑐 β†’ 𝐼 𝑁 c:I\rightarrow N italic_c : italic_I β†’ italic_N , is called a ρ 𝜌 \rho italic_ρ -admissible curve if

( ρ ∘ c ) ⁒ ( t ) = c ~ Λ™ ⁒ ( t ) , 𝜌 𝑐 𝑑 Λ™ ~ 𝑐 𝑑 (\rho\circ c)(t)=\dot{\tilde{c}}(t)\,, ( italic_ρ ∘ italic_c ) ( italic_t ) = Λ™ start_ARG ~ start_ARG italic_c end_ARG end_ARG ( italic_t ) ,

for all t ∈ I 𝑑 𝐼 t\in I italic_t ∈ italic_I . The projection c ~ ~ 𝑐 \tilde{c} ~ start_ARG italic_c end_ARG of a ρ 𝜌 \rho italic_ρ -admissible curve will be called a base curve.