Definition 2.3 .

For p Q 𝑝 𝑄 p\in Q italic_p ∈ italic_Q and η dom ( p ) 𝜂 normal-dom 𝑝 \eta\in\operatorname{dom}(p) italic_η ∈ roman_dom ( italic_p ) we let

p η = p { ρ dom ( p ) : η ρ } . superscript 𝑝 delimited-⟨⟩ 𝜂 𝑝 conditional-set 𝜌 dom 𝑝 𝜂 𝜌 p^{\langle\eta\rangle}=p\restriction\{\rho\in\operatorname{dom}(p)\,:\,\eta% \subseteq\rho\}. italic_p start_POSTSUPERSCRIPT ⟨ italic_η ⟩ end_POSTSUPERSCRIPT = italic_p ↾ { italic_ρ ∈ roman_dom ( italic_p ) : italic_η ⊆ italic_ρ } .

Definition 2.1 (coalgebra, coassociativity)

Let C 𝐶 C italic_C be (generally infinite dimensional) graded vector space. When a coproduct : C C C : 𝐶 tensor-product 𝐶 𝐶 \triangle:C\longrightarrow C\otimes C △ : italic_C ⟶ italic_C ⊗ italic_C is defined on C 𝐶 C italic_C and it is coassociative , i.e.

( 𝟏 ) = ( 𝟏 ) tensor-product 1 tensor-product 1 (\triangle\otimes{\bf 1})\triangle=({\bf 1}\otimes\triangle)\triangle ( △ ⊗ bold_1 ) △ = ( bold_1 ⊗ △ ) △

then C 𝐶 C italic_C is called a coalgebra .

Definition 2.2 (coderivation)

A linear operator 𝔪 : C C : 𝔪 𝐶 𝐶 {\mathfrak{m}}:C\rightarrow C fraktur_m : italic_C → italic_C raising the degree of C 𝐶 C italic_C by one is called coderivation when

𝔪 = ( 𝔪 𝟏 ) + ( 𝟏 𝔪 ) 𝔪 tensor-product 𝔪 1 tensor-product 1 𝔪 \triangle{\mathfrak{m}}=({\mathfrak{m}}\otimes{\bf 1})\triangle+({\bf 1}% \otimes{\mathfrak{m}})\triangle △ fraktur_m = ( fraktur_m ⊗ bold_1 ) △ + ( bold_1 ⊗ fraktur_m ) △

is satisfied. Here, for x , y C 𝑥 𝑦 𝐶 x,y\in C italic_x , italic_y ∈ italic_C , the sign is defined ( 𝟏 𝔪 ) ( x y ) = ( - 1 ) x ( x 𝔪 ( y ) ) tensor-product 1 𝔪 tensor-product 𝑥 𝑦 superscript 1 𝑥 tensor-product 𝑥 𝔪 𝑦 ({\bf 1}\otimes{\mathfrak{m}})(x\otimes y)=(-1)^{x}(x\otimes{\mathfrak{m}}(y)) ( bold_1 ⊗ fraktur_m ) ( italic_x ⊗ italic_y ) = ( - 1 ) start_POSTSUPERSCRIPT italic_x end_POSTSUPERSCRIPT ( italic_x ⊗ fraktur_m ( italic_y ) ) through this operation where the x 𝑥 x italic_x on ( - 1 ) 1 (-1) ( - 1 ) denotes the degree of x 𝑥 x italic_x .

Definition 2.4 ( A subscript 𝐴 A_{\infty} italic_A start_POSTSUBSCRIPT ∞ end_POSTSUBSCRIPT -algebra)

Let {\cal H} caligraphic_H be a graded vector space and C ( ) = k 1 k 𝐶 subscript direct-sum 𝑘 1 superscript tensor-product absent 𝑘 C({\cal H})=\oplus_{k\geq 1}{{\cal H}^{\otimes k}} italic_C ( caligraphic_H ) = ⊕ start_POSTSUBSCRIPT italic_k ≥ 1 end_POSTSUBSCRIPT caligraphic_H start_POSTSUPERSCRIPT ⊗ italic_k end_POSTSUPERSCRIPT be its tensor coalgebra. An A subscript 𝐴 A_{\infty} italic_A start_POSTSUBSCRIPT ∞ end_POSTSUBSCRIPT -algebra is a coalgebra C ( ) 𝐶 C({\cal H}) italic_C ( caligraphic_H ) with a coderivation 𝔪 𝔪 {\mathfrak{m}} fraktur_m which satisfies

( 𝔪 ) 2 = 0 . superscript 𝔪 2 0 ({\mathfrak{m}})^{2}=0\ . ( fraktur_m ) start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT = 0 .

Definition 3.4 .

Let R be a commutative ring. An isotropy ring = D R , 𝒪 subscript superscript R superscript 𝒪 D \cal I=I^{{\mathrm{R},}O^{\prime}}_{\mathrm{D}} caligraphic_I = caligraphic_I start_POSTSUPERSCRIPT roman_R , caligraphic_O start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUPERSCRIPT start_POSTSUBSCRIPT roman_D end_POSTSUBSCRIPT is generated by m o r ( h 𝒪 ) m 𝑜 𝑟 superscript 𝒪 {\mathrm{m}or}(h\cal O^{\prime}) roman_m italic_o italic_r ( italic_h caligraphic_O start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT ) as a free R - mod 𝑅 mod R-\textup{mod} italic_R - mod . Define the multiplication on the generators by

f g = { f g , if codom ( g ) = dom ( f ) 0 , otherwise 𝑓 𝑔 cases 𝑓 𝑔 if codom 𝑔 dom 𝑓 0 otherwise fg=\left\{\begin{array}[]{rl}f\mathbin{\circ}g,&\textup{if\;}\mathrm{codom}({g% })=\mathrm{dom}({f})\\ 0,&\textup{otherwise}\end{array}\right. italic_f italic_g = { start_ARRAY start_ROW start_CELL italic_f ∘ italic_g , end_CELL start_CELL if roman_codom ( italic_g ) = roman_dom ( italic_f ) end_CELL end_ROW start_ROW start_CELL 0 , end_CELL start_CELL otherwise end_CELL end_ROW end_ARRAY

and extend the definition to the rest of the elements of \cal I caligraphic_I by linearity.