ebook img

The variety of $2$-dimensional algebras over an algebraically closed field PDF

0.21 MB·
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview The variety of $2$-dimensional algebras over an algebraically closed field

Thevarietyof2-dimensionalalgebras overan algebraically closedfield 1 IvanKaygorodova,YuryVolkovb aUniversidadeFederaldoABC,CMCC,SantoAndre´,Brazil. bSaintPetersburgstateuniversity,SaintPetersburg,Russia. 7 E-mailaddresses: 1 IvanKaygorodov([email protected]), YuryVolkov([email protected]). 0 2 Abstract.Theworkisdevotedtothevarietyof2-dimensionalalgebrasoveranalgebraicallyclosedfield. Firstly,weclassify n such algebras modulo isomorphism. Then we describe the degenerations and the closures of principal algebra series in the a varietyunderconsideration. Finally, we applyourresultsto obtainanalogousdescriptionsforthe subvarietiesofflexible, and J bicommutativealgebras.Inparticular,wedescriberigidalgebrasandirreduciblecomponentsforthesesubvarieties. 8 2 Keywords:2-dimensionalalgebras,orbitclosure,degeneration,rigidalgebra ] A 1. INTRODUCTION R . Inthispaper,analgebraissimplyavectorspaceoverafieldwithbilinearbinaryoperationthatdoesn’thavetobeassociative. h Algebras of a fixed dimension form a variety with a natural action of a general linear group. The algebraic classification of t a algebras is the classification of isomorphism classes of algebras. In other words it is the classification of orbits under the m action of the generallinear group. There are many classifications for varietiesof algebrasof some fixed dimensionsatisfying [ somepolynomialidentities. Forexample,thereexistalgebraicclassificationsof2-dimensionalpre-Liealgebras,3-dimensional 1 Novikovalgebras,4-dimensionalJordanalgebras,4-dimensionalLeibnizalgebras,and6-dimensionalLiealgebras. v In this paper we classify all 2-dimensional algebras over an algebraically closed field. It is not the first work devoted to 3 this problem, classifications of differenttypeswere made in [1,10,24], but some of them contain inaccuracies, some of them 3 are notwellorganizedandallof themarenotconvenientforourmaingoal, thegeometricdescriptionof thealgebraicvariety 2 of 2-dimensionalalgebras. By this reason, we give a classification that is valid over an algebraically closed field of arbitrary 8 0 characteristic. Inthesamepartofthepaper,wealsodescribeautomorphismgroupsforallalgebrasunderconsideration. Note . thatthereareotherinterestingstructuresonthesetof2-dimensionalalgebras.Oneofthem,thestructureofconservativealgebra 1 definedbyI.Kantor,wasdevelopedin[17,18]. 0 7 In the main part of our paper we developthe geometryof the variety of 2-dimensionalalgebras. Namely, we describe the 1 closures of orbits of some sets with respect to Zariski topology. Firstly, we describe all possible degenerations, i.e. closures : of orbits of one point sets. Degenerations is an interesting subject, which was studied in various papers (see, for example, v i [2–9,11–13,15,16,21–23,25]). Oneoftheproblemsinthisdirectionistodescribealldegenerationsinavarietyofalgebrasof X somefixeddimensionsatisfyingsomesetofidentities. Forexample,thisproblemwassolvedfor2-dimensionalpre-Liealgebras r in[2],for3-dimensionalNovikovalgebrasin[3],for4-dimensionalLiealgebrasin[6],for4-dimensionalZinbielandnilpotent a Leibniz algebras in [16], for nilpotent 5- and 6-dimensional Lie algebras in [11,25], and for nilpotent 5- and 6-dimensional Malcevalgebrasin[15]. Asanapplicationofourresults,onecaneasilyrecovertheresultsof[2]. Anotherinteresting notionconcerningdegenerationsis the so-called levelof an algebra. The algebrasof the first leveland theassociative,LieandJordanalgebrasofthesecondlevelareclassifiedin[19,20]. Inthepapers[7–9],theauthordefinedthe notionofaninfinitelevelanddescribedallanticommutativealgebrasthathaveaninfinitelevelnotgreaterthan3. Thisnotion ismucheasier inthesense thattheinfinitelevelofan algebracanbe easilyexpressedintermsoftheusuallevel. Algebrasof low dimensionplay a specialrole in problemsof such type, because theyhave small levels. The descriptionof degenerations obtainedinthisworkallowstocomputethelevelforall2-dimensionalalgebras. Thenextresultofthispaperisthedescriptionofclosuresoforbitsofsomeprincipalseriesthatappearduringthealgebraic classification.Anotherimportantcharacteristicofavarietyisitspartitiontoirreduciblecomponents.Thenotionofarigidalgebra is closely related to this characteristic, because orbit closures of such algebras form irreducible components. For example, 1 I.KaygorodovwassupportedbyFAPESP14/24519-8. 1 2 irreducible componentsand rigid algebras were classified for low dimensional associative (see [22,23]) and Jordan algebras. Since the varietyof2-dimensionalalgebrasis simplyk8, it isclear thatthereis onlyone irreduciblecomponentandthere are norigidalgebras. Thus,thisproblemisnotrelevantforthevarietyofall2-dimensionalalgebras. Nevertheless,itisrelevantfor subvarieties. Inthelastpartweapplyourresultsaboutthevarietyofall2-dimensionalalgebrastoitssubvarietiesconsistingof flexible and bicommutativealgebras. We describe all degenerationsand closuresof orbits in these varieties. In particular, we classifytheirreduciblecomponentsandrigidalgebras.Ourresultsallowtogetsuchdescriptionsandclassificationsforvarieties of2-dimensionalalgebrasdefinedbyanyidentitieswithoutanyproblems. 2. DEFINITIONS AND NOTATION Duringthepaperwefixanalgebraicallyclosedfieldk,a2-dimensionalk-linearvectorspaceV andabasis{e ,e }ofV. All 1 2 spacesinthispaperareconsideredoverk,andwewritesimplydim,Homand⊗insteadofdimk,Homk and⊗k. Analgebra AisasetwithastructureofavectorspaceandabinaryoperationthatinducesabilinearmapfromA×AtoA. Sincethispaperisdevotedto2-dimensionalalgebras,wegivealldefinitionsandnotationonlyforthiscase,thougheverything inthissectioncanberewrittenforanydimension. ThesetA := Hom(V ⊗V,V) ∼= V∗ ⊗V∗⊗V isavectorspaceofdimension8. Thisspacehasastructureoftheaffine 2 varietyk8.Indeed,anyµ∈A isdeterminedby8structureconstantsck ∈k(i,j,k =1,2)suchthatµ(e ⊗e )=c1e +c2e . 2 ij i j ij 1 ij 2 AsubsetofA isZariski-closedifitcanbedefinedbyasetofpolynomialequationsinthevariablesck. 2 ij ThegenerallineargroupGL(V)actsonA byconjugations: 2 (g∗µ)(x⊗y)=gµ(g−1x⊗g−1y) forx,y ∈V,µ ∈ A andg ∈GL(V). Thus,A isdecomposedintoGL(V)-orbitsthatcorrespondtotheisomorphismclasses 2 2 of2-dimensionalalgebras.Thus,analgebraicclassificationof2-dimensionalalgebrasistheclassificationofGL(V)-orbits. LetO(µ)denotetheorbitofµ ∈ A undertheactionofGL(V)andO(µ)denotetheZariskiclosureofO(µ). LetAandB 2 betwo2-dimensionalalgebrasandµ,λ∈A representAandBrespectively.WesaythatAdegeneratestoBandwriteA→B 2 ifλ∈O(µ). NotethatinthiscasewehaveO(λ)⊂O(µ). Hence,thedefinitionofadegenerationdoesn’tdependonthechoice of µ and λ. If A 6∼= B, then the assertion A → B is called a properdegeneration. We write A 6→ B if λ 6∈ O(µ). Let now A(∗) := {A(α)} beasetof2-dimensionalalgebrasandµ(α) ∈ A representA(α)forα ∈ I. Ifλ ∈ {O(µ(α))} ,then α∈I 2 α∈I wewriteA(∗)→B andsaythatA(∗)degeneratestoB. IntheoppositecasewewriteA(∗)6→B. LetA(∗), B, µ(α) (α ∈ I) andλbe asabove. Letck (i,j,k = 1,2)be thestructureconstantsof λin the basise ,e . If ij 1 2 we constructaj : k∗ → k (i,j = 1,2) and f : k∗ → I such that a1(t)e +a2(t)e and a1(t)e +a2(t)e form a basis of i 1 1 1 2 2 1 2 2 V foranyt ∈ k∗, andthe structure constantsof µ in thisbasis are such polynomialsck(t) ∈ k[t] thatck(0) = ck, then f(t) ij ij ij A(∗) → B. In this case (a1(t)e +a2(t)e ,a1(t)e +a2(t)e ) and f(t) are called a parametrizedbasis and a parametrized 1 1 1 2 2 1 2 2 indexforA(∗) → B respectively. Thecaseofdegenerationbetweentwoalgebrascorrespondstothecase|I| = 1. Inthiscase weneedonlyparametrizedbasis,becausef(t)istheuniqueelementofI foranyt∈k∗. Wetaketheideasforprovingnon-degenerationsfrom[25]. LetQbeasetofpolynomialequationsinvariablesxk (i,j,k= i,j 1,2).SupposethatQsatisfiesthefollowingproperty.Ifxk =ck isasolutiontoallequationsinQ,thenxk =c˜k isasolution i,j ij i,j ij toallequationsinQtoointhefollowingcases: (1) thereareα ,α ∈k∗ suchthatc˜k = αiαjck; (2) thereisα∈1ks2uchthat ij αk ij c˜1 =c1 +α(c1 +c1 )+α2c1 , c˜1 =c1 +αc1 , c˜1 =c1 +αc1 , c˜1 =c1 , 11 11 12 21 22 21 21 22 12 12 22 22 22 c˜2 =c2 +α(c2 +c2 −c1 )+α2(c2 −c1 −c1 )−α3c1 , 11 11 12 21 11 22 12 21 22 c˜2 =c2 +α(c2 −c1 )−α2c1 , c˜2 =c2 +α(c2 −c1 )−α2c1 , c˜2 =c2 −αc1 . 21 21 22 21 22 12 12 22 12 22 22 22 22 LetR ⊂ A bea setofallalgebrastructureswhosestructureconstantssatisfy allequationsinQ. We willcallsucha setR a 2 closedupperinvariantset. Let{A(α)} beasetof2-dimensionalalgebrassuchthatA(α)canberepresentedbyastructure α∈I fromRforanyα∈I. LetBbea2-dimensionalalgebrarepresentedbythestructureλ∈A . IfO(λ)∩R =∅,thenA(∗)6→B. 2 InthiscasewecallRaseparatingsetforA(∗)6→B. 3 Let us recall two more tools for proving degenerations and non-degenerations. Firstly, if A → B, then dimAut(A) < dimAut(B). Note that if A(∗) 6→ B, then either dimAut(A(α)) = dimAut(B) for infinitely many α ∈ I or dimAut(A(α))<dimAut(B)forsomeα∈I,butitispossiblethatdimAut(A(α))≥dimAut(B)forallα∈I. Notealso thatdimAut(A)=dimDer(A). Secondly,ifA→C andC →BthenA→B. IfthereisnoC suchthatA→CandC →B areproperdegenerations,thentheassertionA→B iscalledaprimarydegeneration.IftherearenoC andDsuchthatC →A, B → D, C 6→ D andoneoftheassertionsC → AandB → D isaproperdegeneration,thentheassertionA 6→ B iscalled aprimarynon-degeneration.Itsufficestoproveonlyprimarydegenerationsandnon-degenerationstodescribedegenerationsin thevarietyunderconsideration.Notealsothatanyalgebradegeneratestothealgebrawithzeromultiplication. 3. ALGEBRAIC CLASSIFICATION Thefirstofouraimsistoclassifyall2-dimensionalalgebrasoverkmoduloisomorphism. Ourclassificationisbasedonthe followinglemma. Lemma 1. Let A be a 2-dimensionalalgebra. Then there exists a non-zero element x ∈ A such that x and x2 are linearly dependent. Proof. Ifc1 =0,wecantakex=e . Supposethatc1 6=0. Letusdefinex =e +te fort∈k. Since 22 2 22 t 1 2 x2 =(c1 +(c1 +c1 )t+c1 t2)e +(c2 +(c2 +c2 )t+c2 t2)e , t 11 12 21 22 1 11 12 21 22 2 x andx2 arelinearlydependentiff t t c1 +(c1 +c1 )t+c1 t2 c2 +(c2 +c2 )t+c2 t2 0= 11 12 21 22 11 12 21 22 =c1 t3+At2+Bt+C, (cid:12) 1 t (cid:12) 22 (cid:12) (cid:12) whereA,B,C ∈ kare(cid:12)(cid:12) someelementsdependingonlyonthestructureconstants.(cid:12)(cid:12)Sincec1 6= 0andkisalgebraicallyclosed, 22 thereexiststsuchthatx andx2 arelinearlydependent. t t ✷ Notethatifx ∈ Aandx2 arelinearlydependent,theneitherx2 = 0orx = αeforsomeα ∈ k∗ andsomee ∈ Asuchthat e2 =e. Ifx2 =0,thenxiscalleda2-nilelement.Anelementesuchthate2 =eiscalledanidempotent. Corollary2. Any2-dimensionalk-algebrabelongstooneofthefollowingdisjointclasses: A. algebrasthatdon’thavenonzeroidempotentsandhaveaunique1-dimensionalsubspaceof2-nilelements; B. algebrasthatdon’thavenonzeroidempotentsandhavetwolinearlyindependent2-nilelements; C. algebrasthathaveauniquenonzeroidempotentanddon’thavenonzero2-nilelements; D. algebrasthathaveauniquenonzeroidempotentandanonzero2-nilelement; E. algebrasthathavetwodifferentnonzeroidempotents. Proof. The fact thatthe classes are disjointis obvious. The fact thatany 2-dimensionalalgebra belongsto one of the classes followseasilyfromLemma1andtheremarkafterit. ✷ To give the classification of 2-dimensionalalgebras we have to introduce some notation. Let us consider the action of the cyclicgroupC = hρ | ρ2ionkdefinedbytheequalityρα = −αforα ∈ k. Letusfixsomesetofrepresentativesoforbits 2 underthisactionanddenoteitbyk≥0. Forexample,ifk = C, thenonecantakeC≥0 = {α ∈ C | Re(α) > 0}∪{α ∈ C | Re(α)=0,Im(α)≥0}. LetusalsoconsidertheactionofC onk2definedbytheequalityρ(α,β)=(1−α+β,β)for(α,β)∈k2. Letusfixsome 2 setofrepresentativesoforbitsunderthisactionanddenoteitbyU. LetusalsodefineT ={(α,β)∈k2 |α+β =1}. Given(α,β,γ,δ) ∈ k4,wedefineD(α,β,γ,δ) = (α+γ)(β+δ)−1. WedefineC (α,β,γ,δ) = (β,δ),C (α,β,γ,δ) = 1 2 (γ,α),andC (α,β,γ,δ) = βγ−(α−1)(δ−1),αδ−(β−1)(γ−1) for(α,β,γ,δ)suchthatD(α,β,γ,δ) 6= 0. Letusconsiderthe 3 D(α,β,γ,δ) D(α,β,γ,δ) (cid:16) (cid:17) set C (Γ),C (Γ),C (Γ) |Γ∈k4,D(Γ)6=0,C (Γ),C (Γ)6∈T ⊂ (k2)3. OnecanshowthatthesymmetricgroupS acts 1 2 3 1 2 3 ont(cid:8)h(cid:0)issetbytheequalityσ(cid:1) C (Γ),C (Γ),C (Γ) = C (Γ),C(cid:9) (Γ),C (Γ) forσ ∈S . Notethatthereexistsa 1 2 3 σ−1(1) σ−1(2) σ−1(3) 3 (cid:0) (cid:1) (cid:0) (cid:1) 4 setofrepresentativesoforbitsunderthisactionV˜ suchthatif(C ,C ,C ) ∈ V˜ andC 6= C ,thenC 6= C ,C . Letusfixsuch 1 2 3 1 2 3 1 2 V˜ anddefine V={Γ∈k4 |D(Γ)6=0;C (Γ),C (Γ)6∈T, C (Γ),C (Γ),C (Γ) ∈V˜}. 1 2 1 2 3 ForΓ∈V,wealsodefineC(Γ)={C (Γ),C (Γ),C (Γ)}⊂k2. (cid:0) (cid:1) 1 2 3 Let us consider the action of the cyclic group C on k∗ \{1} defined by the equality ρα = α−1 for α ∈ k∗ \{1}. Let 2 us fix some set of representativesof orbitsunderthis actionand denoteit by k∗ . For example, if k = C, then onecan take >1 C∗ ={α∈C∗ ||α|>1}∪{α∈C∗ ||α|=1,0<arg(α)≤π}. For(α,β,γ)∈k2×k∗ wedefine >1 >1 β 1−β C(α,β,γ)= αγ,(1−α)γ , , ⊂k2. (cid:26) (cid:18)γ γ (cid:19)(cid:27) (cid:0) (cid:1) LetF ⊂A bethesetformedbythefollowingalgebrastructuresonV: 2 A (α),α∈k e e =e +e , e e =αe , e e =(1−α)e , e e =0 1 1 1 1 2 1 2 2 2 1 2 2 2 A e e =e , e e =e , e e =−e , e e =0 2 1 1 2 1 2 2 2 1 2 2 2 A e e =e , e e =0, e e =0, e e =0 3 1 1 2 1 2 2 1 2 2 A4(α),α∈k≥0 e1e1 =αe1+e2, e1e2 =e1+αe2, e2e1 =−e1, e2e2 =0 B (α),α∈k e e =0, e e =(1−α)e +e , e e =αe −e , e e =0 1 1 1 1 2 1 2 2 1 1 2 2 2 B (α),α∈k e e =0, e e =(1−α)e , e e =αe , e e =0 2 1 1 1 2 1 2 1 1 2 2 B e e =0, e e =e , e e =−e , e e =0 3 1 1 1 2 2 2 1 2 2 2 C(α,β),(α,β)∈k×k≥0 e1e1 =e2, e1e2 =(1−α)e1+βe2, e2e1 =αe1−βe2, e2e2 =e2 D (α,β),(α,β)∈U e e =e , e e =(1−α)e +βe , e e =αe −βe , e e =0 1 1 1 1 1 2 1 2 2 1 1 2 2 2 D (α,β),(α,β)∈k2\T e e =e , e e =αe , e e =βe , e e =0 2 1 1 1 1 2 2 2 1 2 2 2 D (α,β),(α,β)∈k2\T e e =e , e e =e +αe , e e =−e +βe , e e =0 3 1 1 1 1 2 1 2 2 1 1 2 2 2 E (α,β,γ,δ),(α,β,γ,δ)∈V e e =e , e e =αe +βe , e e =γe +δe , e e =e 1 1 1 1 1 2 1 2 2 1 1 2 2 2 2 E (α,β,γ), 2 e e =e , e e =(1−α)e +βe , e e =αe +γe , e e =e (α,β,γ)∈k3\k×T 1 1 1 1 2 1 2 2 1 1 2 2 2 2 E (α,β,γ), 3 e e =e , e e =(1−α)γe + βe , e e =αγe + 1−βe , e e =e (α,β,γ)∈k2×k∗ 1 1 1 1 2 1 γ 2 2 1 1 γ 2 2 2 2 >1 E e e =e , e e =e +e , e e =0, e e =e 4 1 1 1 1 2 1 2 2 1 2 2 2 E (α),α∈k e e =e , e e =(1−α)e +αe , e e =αe +(1−α)e , e e =e 5 1 1 1 1 2 1 2 2 1 1 2 2 2 2 Thissectionisdevotedtotheproofofthefollowingtheoremthatgivesanalgebraicclassificationof2-dimensionalalgebras overk. Theorem3. Anynon-trivial2-dimensionalk-algebracanberepresentedbyauniquestructurefromF. Inotherwords,Theorem3statesthatA = O(µ)∪{k2}andthat,ifµ,λ∈Faredifferentstructures,thenO(µ)∩O(λ) = 2 µS∈F ∅. Whenever an algebra named A appears in this section, we suppose that it is represented by some structure from A with 2 structureconstantsck (i,j,k =1,2). TherestofthissectionisdevotedtotheproofofTheorem3. AccordingtoCorollary2,it ij sufficestoconsidereachoftheclassesA–Eseparately. Lemma4. IfAbelongstotheclassA,thenitcanberepresentedbyauniquestructurefromtheset (1) {A1(α)}α∈k∪{A2}∪{A3}∪{A4(α)}α∈k≥0. Proof. Let us represent the algebra A by a structure such that e e = 0. It is easy to see that A belongs to the class A iff 2 2 x =e +te andx2 arelinearlyindependentforanyt∈k. Since t 1 2 t x2 =(c1 +(c1 +c1 )t)e +(c2 +(c2 +c2 )t)e , t 11 12 21 1 11 12 21 2 x andx2 arelinearlyindependentiff t t c1 +(c1 +c1 )t c2 +(c2 +c2 )t 06= 11 12 21 11 12 21 =(c1 +c1 )t2+(c1 −c2 −c2 )t−c2 . (cid:12) 1 t (cid:12) 12 21 11 12 21 11 (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) 5 Sincebyourassumptionx andx2arelinearlyindependentforanyt∈k,wehavec1 +c1 =0,c1 =c2 +c2 ,andc2 6=0. t t 12 21 11 12 21 11 Nowwehavefourcases: • c112 =0,c111 6=0. Consideringthebasis ce11 , (cc2111e)22 ofV,onecancheckthatAcanberepresentedbyA1 cc1221 . 11 11 (cid:16) 11(cid:17) • c1 =0,c2 =−c2 6=0. Consideringthebasis e1 , c211e2 ofV,onecancheckthatAcanberepresentedbyA . 12 12 21 c2 (c2 )2 2 12 12 • c1 =c2 =c2 =0. Consideringthebasise ,c2 e ofV,onecancheckthatAcanberepresentedbyA . 12 12 21 1 11 2 3 • c112 6= 0. Leta ∈ k∗ besuchthatc211c112a2 = 1andc111a ∈ k≥0. Consideringthebasisa e1− cc2211e2 , ce12 ofV,one (cid:16) 12 (cid:17) 12 cancheckthatAcanberepresentedbyA (c1 a). 4 11 It remains to prove that any two different structures from the set (1) represent non-isomorphicalgebras. Firstly, note that dim(A )2 = dim(A )2 = 1whiledim(A (α))2 = dim(A (α))2 = 2foranyα ∈ k. WehavealsoA 6∼= A ,becauseA 2 3 1 4 2 3 3 hasanonzeroannihilator. SupposethatAisrepresentedbythestructureA (α)forsomeα∈k. Thenthereexistsx∈Asuchthatx2 =0,xA+Ax⊂ 1 hxi, and αxy = (1−α)yx for any y ∈ A. Such an element doesn’t exist in A (β) for any β ∈ k and in A (β) for any 4 1 β ∈k\{α}. SupposethatA is representedby the structure A4(α) forsome α ∈ k≥0. Supposethatthe structure constantsof A in the basisE1, E2 equalthestructureconstantsofA4(β)forsomeβ ∈ k≥0. SinceE2E2 = 0andE2E1 = −E1, itiseasytosee thatE =e andE = ae forsomea∈ k∗. ThenweobtainfromtheequalityE E =βE +E thata =±1andβ =±α. 2 2 1 1 1 1 1 2 Sinceα,β ∈k≥0,wehaveβ =α. ✷ Lemma5. IfAbelongstotheclassB,thenitcanberepresentedbyauniquestructurefromtheset (2) {B1(α)}α∈k∪{B2(α)}α∈k∪{B3}. Proof. LetusrepresentthealgebraAbyastructuresuchthate e = e e = 0. Fors,t ∈ k, letusdefinex = se +te . 1 1 2 2 s,t 1 2 Supposethatthereare s,t ∈ k∗ suchthat0 = x2 = st(e e +e e ). Thene e +e e = 0 andA isanticommutative. It s,t 1 2 2 1 1 2 2 1 iseasytoseethatany2-dimensionalanticommutativealgebraeitherhasthetrivialmultiplicationorcanberepresentedbyB 3 (notethatbyourdefinitionAisanticommutativeiffx2 =0foranyx∈A). Supposenowthatx2 6= 0 foranys,t ∈ k∗. SinceAdoesn’thaveidempotents,x andx2 arelinearlyindependentfor s,t s,t s,t s,t ∈ k∗. Itiseasytocheckthatx andx2 arelinearlydependentfors = c2 +c2 ,t = c1 +c1 . Hence,c1 +c1 = 0 s,t s,t 12 21 12 21 12 21 or c2 +c2 = 0. Without loss of generality we may assume that c2 +c2 = 0. Since A is not anticommutative, we have 12 21 12 21 c1 +c1 6=0inthiscase. 12 21 If c2 6= 0, then, considering the basis e1 , e2 of V, one can check that A can be represented by B c121 . If 12 c2 c1 +c1 1 c1 +c1 12 12 21 (cid:16) 12 21(cid:17) c2 =0,then,consideringthebasise , e2 ofV,onecancheckthatAcanberepresentedbyB c121 . 12 1 c1 +c1 2 c1 +c1 It remains to prove that any two diffe1r2ent21structures from the set (2) represent non-isomorphical(cid:16)ge1b2ras2.1(cid:17)Since B is anti- 3 commutative,it isnotisomorphicto otheralgebrasfrom(2). Notealso thatdim(B (α))2 = 2 > 1 = dim(B (β))2 forany 1 2 α,β ∈k. SupposethatAisrepresentedbythestructureB (α)forsomeα∈kandi=1,2. SupposethatthestructureconstantsofA i inthebasisE ,E equalthestructureconstantsofB (β)forsomeβ ∈k. SinceE E =E E =0,wehaveeitherE =ae , 1 2 i 1 1 2 2 1 1 E = be orE = ae ,E = be forsomea,b ∈ k∗. SinceE E +E E = E ,wehaveE = ae andE = e . Thenwe 2 2 1 2 2 1 1 2 2 1 1 1 1 2 2 getβ =αfromtheequalityE E =(1−β)E +(2−i)E . 1 2 1 2 ✷ Lemma6. IfAbelongstotheclassC,thenitcanberepresentedbyC(α,β)forauniquepair(α,β)∈k×k≥0. Proof. Let us represent the algebra A by a structure such that e e = e . It is easy to see that A belongs to the class C iff 2 2 2 x =e +te andx2 arelinearlyindependentforanyt∈k. Since t 1 2 t x2 =(c1 +(c1 +c1 )t)e +(c2 +(c2 +c2 )t+t2)e , t 11 12 21 1 11 12 21 2 6 x andx2 arelinearlyindependentiff t t c1 +(c1 +c1 )t c2 +(c2 +c2 )t+t2 06= 11 12 21 11 12 21 =(c1 +c1 −1)t2+(c1 −c2 −c2 )t−c2 . (cid:12) 1 t (cid:12) 12 21 11 12 21 11 (cid:12) (cid:12) Sincebyourassu(cid:12)(cid:12)mptionxtandx2t arelinearlyindependentforan(cid:12)(cid:12)yt∈k,wehavec112+c121 =1,c111 =c212+c221,andc211 6=0. Letabesuchanelementofk∗ thatc211a2 =1anda(c212−c111c121)∈k≥0. Consideringthebasisa(e1−c111e2),e2ofV,one cancheckthatAcanberepresentedbyC c1 ,a(c2 −c1 c1 ) . 21 12 11 21 SupposethatAisrepresentedbythestr(cid:0)uctureC(α,β)forso(cid:1)mepair(α,β)∈k×k≥0. Supposethatthestructureconstants ofAinthebasisE1, E2 equalthestructureconstantsofC(γ,δ)forsome(γ,δ) ∈ k×k≥0. SinceE2E2 = E2 andC(α,β) hasauniqueidempotent,wehaveE =e . WegetE =±e fromtheequalityE E = E . Thenγ = αandδ = ±β. Since 2 2 1 1 1 1 2 β,δ ∈k≥0,wehave(γ,δ)=(α,β). ✷ Lemma7. IfAbelongstotheclassD,thenitcanberepresentedbyauniquestructurefromtheset (3) {D1(α,β)}(α,β)∈U∪{D2(α,β)}(α,β)∈k2\T∪{D3(α,β)}(α,β)∈k2\T. Proof. LetusrepresentthealgebraAbyastructuresuchthate e =e ande e =0. 1 1 1 2 2 Letusconsiderthefollowingcases: • c1 +c1 6=0. Ifc2 +c2 6=0,thenonecancheckthat 1 e + c212+c221−1e isanidempotentthatisnotequal 12 21 12 21 c2 +c2 1 c1 +c1 2 12 21 (cid:16) 12 21 (cid:17) toe . Thus,c2 +c2 = 0. If c112 ,c2 ∈ U,then,consideringthebasise , e2 ofV,onecancheckthatA 1 12 21 c1 +c1 12 1 c1 +c1 (cid:16) 12 21 (cid:17) 12 21 canberepresentedbyD1 c1c+112c1 ,c212 . If c1c+112c1 ,c212 6∈ U, then c1c+121c1 +c212,c212 ∈ Uand,consideringthe (cid:16) 12 21 (cid:17) (cid:16) 12 21 (cid:17) (cid:16) 12 21 (cid:17) basise ,e − e2 ofV,onecancheckthatAcanberepresentedbyD c121 +c2 ,c2 . 1 1 c1 +c1 1 c1 +c1 12 12 12 21 (cid:16) 12 21 (cid:17) • c1 = −c1 6= 0. Consideringthebasise , e2 ofV,onecancheckthatAcanberepresentedbyD (c2 ,c2 ). Since 12 21 1 c1 3 12 21 12 e +e isnotidempotent,(c2 ,c2 )6∈T. 1 2 12 21 • c1 =c1 6=0.ThenonecancheckthatAisrepresentedbyD (c2 ,c2 ).Sincee +e isnotidempotent,(c2 ,c2 )6∈T. 12 21 2 12 21 1 2 12 21 Itremainstoprovethatanytwodifferentstructuresfromtheset(3)representnon-isomorphicalgebras. Suppose that A is represented by the structure D (α,β) for some pair (α,β) ∈ U. Note that e e = (e − e )2 = 0 1 2 2 1 2 in D (α,β) while the structures D (γ,δ) and D (γ,δ) have a unique 1-dimensionalsubspace of 2-nilelements for any pair 1 2 3 (γ,δ)∈k2. SupposenowthatthestructureconstantsofAinthebasisE ,E equalthestructureconstantsofD (γ,δ)forsome 1 2 1 pair(γ,δ)∈U. SinceE isanidempotentandD (α,β)hasauniqueidempotent,wehaveE =e . SinceE E =0,wehave 1 1 1 1 2 2 eitherE = ae orE = a(e −e )forsomea ∈ k∗. Weobtaina = 1inbothcasesfromtheequalityE E +E E = E . 2 2 2 1 2 1 2 2 1 1 Thenwehaveδ =β andeitherγ =αorγ =1−α+β. Since(α,β),(γ,δ)∈U,wehave(γ,δ)=(α,β). SupposethatAisrepresentedbythestructureD (α,β)forsomepair(α,β) ∈ k2. NotethatAhasanelementxsuchthat 2 x2 = 0 and xA+Ax ⊂ hxi while D (γ,δ) doesn’t have such an element for any pair (γ,δ) ∈ k2. Suppose now that the 3 structureconstantsofAinthebasisE ,E equalthestructureconstantsofD (γ,δ)forsomepair(γ,δ) ∈ k2. SinceE isan 1 2 2 1 idempotentandD (α,β)hasauniqueidempotent,wehaveE = e . SinceE E = 0,wehaveE = ae forsomea ∈ k∗. 2 1 1 2 2 2 2 Thenitiseasytoseethat(γ,δ)=(α,β). Finally, suppose that A is represented by the structure D (α,β) for some pair (α,β) ∈ k2. Suppose that the structure 3 constantsofAinthebasisE ,E equalthestructureconstantsofD (γ,δ)forsomepair(γ,δ)∈k2. SinceE isanidempotent 1 2 3 1 andD (α,β)hasauniqueidempotent,wehaveE =e . SinceE E =0,wehaveE =ae forsomea∈k∗. Thenitiseasy 3 1 1 2 2 2 2 toseethata=1and(γ,δ)=(α,β). ✷ AsaconsequenceoftheproofsofLemmas4–7wecandescribetheautomorphismgroupsofthealgebrasoftheclassesA–D. Corollary8. 1.Aut(A (α))∼=Aut(A )isisomorphictotheadditivegroupofk. 1 2 a 0 2.Aut(A )isisomorphictothesubgroupofGL (k)formedbymatricesoftheform ,wherea∈k∗ andb∈k. 3 2 (cid:18) b a2 (cid:19) 7 3.Aut(A (α))∼=C ifα=0andchark6=2;Aut(A (α))istrivialifeitherα∈k∗ orα=0andchark=2. 4 2 4 4.Aut(B (α))istrivial;Aut(B (α))∼=k∗. 1 2 1 0 5.Aut(B )isisomorphictothesubgroupofGL (k)formedbymatricesoftheform ,wherea∈k∗ andb∈k. 3 2 (cid:18) b a (cid:19) 6.Aut(C(α,β))∼=C ifβ =0andchark6=2;Aut(C(α,β))istrivialifeitherα∈k∗ orα=0andchark=2. 2 7.Aut(D (α,β))∼=C ifβ =2α−1andAut(C(α,β))istrivialifβ 6=2α−1. 1 2 8.Aut(D (α,β))∼=k∗ andAut(D (α,β))istrivialifα+β 6=1. 2 3 Inparticular, dimAut(A (α))=dimAut(B (α))=dimAut(C (α,β))=dimAut(D (α,β))=dimAut(D (α,β))=0, 4 1 1 1 3 dimAut(A (α))=dimAut(A )=dimAut(B (α))=dimAut(D (α,β))=1anddimAut(A )=dimAut(B )=2. 1 2 2 2 3 3 Proof. 1–3. AnystructureoftheclassAhasauniquesubspaceof2-nilelementsgeneratedbye . Thus,anyautomorphismof 2 suchanalgebrasendse ande toae +be andce respectively,wherea,c∈k∗ andb∈k. Itiseasytocheckthata=c=1 1 2 1 2 2 forA (α)andA ;c=a2forA ;a=±1,b=0andc=1forA (0);anda=c=1,b=0forA (α)ifα6=0. 1 2 3 4 4 4. ItfollowsfromtheproofofLemma5thatanyautomorphismofthealgebraB (α), wherei ∈ {1,2},sendse ande to i 1 2 ae ande respectivelyforsomea∈k∗. Itiseasytoseethata=1fori=1andacanbearbitraryfori=2. 1 2 5. Since B (V,V) is generatedby e , any automorphismof B sends e and e to ae +be and ce respectively,where 3 2 3 1 2 1 2 2 a,c∈k∗ andb∈k. Itiseasytoseethatsuchamapisanautomorphismiffa=1. 6. It follows from the proof of Lemma 6 that any automorphism of the algebra C(α,β) sends e and e to ±e and e 1 2 1 2 respectively. It is easy to see that the map that sends e and e to −e and e respectively is an automorphismiff β = 0 or 1 2 1 2 chark=2. 7. ItfollowsfromtheproofofLemma7thatanyautomorphismofthealgebraD (α,β)sendse toe andsendse eitherto 1 1 1 2 e ore −e . Itiseasytoseethatthemapthatsendse ande toe ande −e respectivelyisanautomorphismiffβ =2α−1. 2 1 2 1 2 1 1 2 8. It follows from the proof of Lemma 7 that any automorphism of the algebra D (α,β) sends e and e to e and ae 2 1 2 1 2 respectivelyforsomea∈k∗. ItfollowsfromthesameproofthatanyautomorphismofthealgebraD (α,β)istrivial. 3 ✷ WewillfinishtheproofofTheorem3inthenextsectiondevotedtothealgebrasoftheclassE. 4. ALGEBRAS OF THECLASSE InthissectionweconsiderthealgebrasoftheclassE. ItisclearthatsuchanalgebraisisomorphictoE (Γ)forsomeΓ∈k4. 1 Firstly,wedescribeisomorphismsinsidethissetand,thus,finishtheproofofTheorem3. Lemma9. E (Γ )∼=E (Γ )iffoneofthefollowingconditionsholds: 1 1 1 2 • Γ =Γ ; 1 2 • C (Γ )=C (Γ )andC (Γ )=C (Γ ); 1 1 2 2 2 1 1 2 • C (Γ ),C (Γ ),C (Γ ),C (Γ )∈T,C (Γ )6=C (Γ ),C (Γ )6=C (Γ ); 1 1 1 2 2 1 2 2 1 1 2 1 1 2 2 2 • C (Γ ),C (Γ )6∈T,D(Γ )6=0,andthereissomeσ ∈S suchthatC (Γ )=C (Γ )fori∈{1,2,3}. 1 2 2 2 2 3 i 1 σ(i) 2 Proof. Supposethatg ∈GL(V)issuchthatg∗E (Γ )=E (Γ ). Thenge andge aretwolinearlyindependentidempotents 1 1 1 2 1 2 ofE (Γ ). Letusdescribeallnonzeroidempotentsofthisalgebra. LetΓ = (α,β,γ,δ)andu = xe +ye besomeelement 1 2 2 1 2 ofV. ThenE (Γ )(u,u)=uiffx=x2+(α+γ)xyandy =(β+δ)xy+y2.Letusconsiderthefollowingcases: 1 2 • C (Γ )=C (Γ )∈T. InthiscaseE (Γ )(u,u)=uiffeitherx=y =0orx+y =1. Thus,ge =ae +(1−a)e 1 2 2 2 1 2 1 1 2 andge =be +(1−b)e fortwodifferenta,b∈k. OnecancheckthatinthiscaseΓ =Γ . 2 1 2 1 2 • C (Γ ),C (Γ ) ∈ T, C (Γ ) 6= C (Γ ), i.e. α + γ = β + δ = 1, α 6= δ. As in the previous case, we have 1 2 2 2 1 2 2 2 ge = ae +(1−a)e andge = be +(1−b)e fortwodifferenta,b ∈ k. Thus,E (Γ ) ∼= E (Γ )iffthereexist 1 1 2 2 1 2 1 1 1 2 a,b∈k,a6=bsuchthat Γ =((1−b)α+bδ,aβ+(1−a)γ,bβ+(1−b)γ,(1−a)α+aδ). 1 ItiseasytoseethatsuchaandbexistiffC (Γ ),C (Γ )∈TandC (Γ )6=C (Γ ). 1 1 2 1 1 1 2 1 8 • oneofthefollowingthreeconditionsholds: 1. C (Γ )∈T,C (Γ )6∈T; 1 2 2 2 2. C (Γ )6∈T,C (Γ )∈T; 1 2 2 2 3. C (Γ ),C (Γ )6∈T,D(Γ )=0. 1 2 2 2 2 Itiseasytoseethat,inallofthesecases,E (Γ )(u,u) = uonlyforu ∈ {0,e ,e }. Thus,eitherge = e ,ge = e 1 2 1 2 1 1 2 2 andΓ =Γ orge =e ,ge =e ,C (Γ )=C (Γ )andC (Γ )=C (Γ ). 1 2 1 2 2 1 1 1 2 2 2 1 1 2 • C (Γ ),C (Γ ) 6∈ T, D(Γ ) 6= 0. In this case, E (Γ )(u,u) = u for u ∈ {0,e ,e ,e }, where e = α+γ−1e + 1 2 2 2 2 1 2 1 2 3 3 D(Γ2) 1 β+δ−1e . Thus, there is σ ∈ S such that ge = e for i ∈ {1,2}. Then direct calculations show that C (Γ ) = D(Γ2) 2 3 i σ(i) i 1 C (Γ )fori∈{1,2,3}. σ(i) 2 ✷ NowwecanfinishtheproofofTheorem3. ProofofTheorem3. ByCorollary2,thealgebraAbelongstooneoftheclassesA–EandtheclasscontainingAisunique. If AbelongstooneoftheclassesA–D,thenthestatementofthetheoremfollowsfromLemmas4–7. SupposethatAbelongstotheclassE. ThenAcanberepresentedbyE (Γ)forsomeΓ=(α,β,γ,δ)∈k4. Nowwehave 1 • ifC (Γ)=C (Γ)∈T,thenE (Γ)=E (β); 1 2 1 4 • ifC (Γ),C (Γ)∈TandC (Γ)6=C (Γ),thenE (Γ)∼=E (1,1,0,0)=E byLemma9; 1 2 1 2 1 1 4 • ifC (Γ)∈TandC (Γ)6∈T,thenE (Γ)∼=E (1−β,γ,β,α)=E (β,γ,α)byLemma9; 1 2 1 1 2 • ifC (Γ)6∈TandC (Γ)∈T,thenE (Γ)=E (γ,β,δ); 1 2 1 2 • ifC (Γ),C (Γ)6∈T,D(Γ)=0andα+γ ∈k∗ ,thenE (Γ)=E γ(β+δ),β(α+γ),α+γ ; 1 2 >1 1 3 • ifC (Γ),C (Γ) 6∈ T,D(Γ) = 0andα+γ 6∈ k∗ ,thenE (Γ) ∼= E(cid:0)(δ,γ,β,α) = E β(α+γ)(cid:1),γ(β+δ),β+δ by 1 2 >1 1 1 3 Lemma9; (cid:0) (cid:1) • ifC (Γ),C (Γ)6∈TandD(Γ)6=0,thenthereisauniqueσ ∈S suchthatσ−1(C (Γ),C (Γ),C (Γ))∈V˜ andwehave 1 2 3 1 2 3 E1(Γ)∼=E1(Γ′)byLemma9,whereΓ′ ∈VissuchthatCi(Γ′)=Cσ(i)(Γ)fori∈{1,2,3}. ByLemma9,thestructuresfromtheset {E1(Γ)}Γ∈V∪{E2(α,β,γ)}(α,β,γ)∈k3\k×T∪{E3(α,β,γ)}(α,β,γ)∈k2×k∗ ∪{E4}∪{E5(α)}α∈k >1 arepairwisenon-isomorphic. ✷ AsaconsequenceoftheproofofLemma9wecandescribetheautomorphismgroupsofalgebrasoftheclassE. Corollary 10. 1. For Γ ∈ V, Aut(E (Γ)) is trivial if C (Γ) 6= C (Γ), Aut(E (Γ)) ∼= C if C (Γ) = C (Γ) 6= (−1,−1), 1 1 2 1 2 1 2 Aut(E (−1,−1,−1,−1))∼=S ifchark6=3. 1 3 2.Aut(E )andAut(E (α,β,γ))aretrivialfor(α,β,γ)∈k3\k×T. 4 2 3.For(α,β,γ)∈k2×k∗ ,Aut(E (α,β,γ))∼=C ifγ =−1andα=β andAut(E (α,β,γ))istrivialotherwise. >1 3 2 3 a b 4.Aut(E (α))isisomorphictothesubgroupofGL (k)formedbymatricesoftheform ,wherea,b∈k, 5 2 (cid:18) 1−a 1−b (cid:19) a6=b. In particular, we have dimAut(E (Γ)) = dimAut(E (α,β,γ)) = dimAut(E (α,β,γ)) = dimAut(E ) = 0 and 1 2 3 4 dimAut(E (α))=2. 5 Proof. 1.AnyautomorphismofE (Γ)hastosende ande toe ande respectivelyforsomeσ ∈S ,wheree isdefined 1 1 2 σ(1) σ(2) 3 3 intheproofofLemma9. SuchamapisanautomorphismiffC (Γ)=C (Γ)fori=1,2.IfC (Γ)6=C (Γ),thenwehavealso i σ(i) 1 2 C (Γ) 6= C (Γ),C (Γ)and,hence,onlyidenticalelementofS determinesanautomorphism. IfC (Γ) = C (Γ) 6= (−1,−1), 3 1 2 3 1 2 then one can check that C (Γ) 6= C (Γ),C (Γ) and, hence, only identical element of S and the element that swaps e and 3 1 2 3 1 e determine automorphisms. If C (Γ) = C (Γ) = (−1,−1), then any σ ∈ S determines an automorphism. Note that 2 1 2 3 (−1,−1,−1,−1)∈Viffchark6=3. 2. FollowsdirectlyfromtheproofofLemma9. 9 3. ItfollowsfromtheproofofLemma9thatanautomorphismofE (α,β,γ)iseitheristrivialorswapse ande . Thelast 3 1 2 mentionedmapisanautomorphismiffγ =−1andα=β. 4. ItfollowsdirectlyfromtheproofofLemma9thatautomorphismsofE (α)areexactlythelinearmapsthatsende and 5 1 e toae +(1−a)e andbe +(1−b)e fortwodifferenta,b∈k. 2 1 2 1 2 ✷ NowwearegoingtodiscusssomefactsaboutdegenerationsoftheformA→B,whereAisanalgebraoftheclassE. First ofall,letusprovethefollowinglemma. Lemma11. 1.ForanyΓ∈Vand(β,γ)∈C(Γ)thereexistsadegenerationE (Γ)→D (β,γ). 1 2 2.Forany(α,β,γ)∈k3\k×TthereexistsadegenerationE (α,β,γ)→D (β,γ). 2 2 3.Forany(α,δ,ǫ)∈k2×k∗ and(β,γ)∈C(α,δ,ǫ)thereexistsadegenerationE (α,δ,ǫ)→D (β,γ). >1 3 2 Proof. TheparametrizedbasisEt = e ,Et = te givesthedegenerationE (Γ) → D C (Γ) foranyΓ ∈ k4. IfΓ ∈ Vand 1 1 2 2 1 2 1 1≤i≤3,then,byLemma9,thereexistsΓ′ ∈k4 suchthatE1(Γ)∼=E1(Γ′)andCi(Γ)(cid:0)=C1(Γ(cid:1)′). Hence,E1(Γ)∼=E1(Γ′)→ D C (Γ) . WealsohaveE (α,β,γ)=E (1−α,β,α,γ)→D (β,γ)for(α,β,γ)∈k3\k×Tand 2 i 2 1 2 (cid:0) (cid:1) δ 1−δ 1−δ δ δ 1−δ E (α,δ,ǫ)=E (1−α)ǫ, ,αǫ, ∼=E ,αǫ, ,(1−α)ǫ →D αǫ,(1−α)ǫ) ,D , 3 1(cid:18) ǫ ǫ (cid:19) 1(cid:18) ǫ ǫ (cid:19) 2 2(cid:18)ǫ ǫ (cid:19) (cid:0) (cid:1) for(α,δ,ǫ)∈k2×k∗ . >1 ✷ Letusdefine,forΓ=(α,β,γ,δ)∈k4,thefollowingsubsetofA : 2 c1 =0, c1 =γc2 , c1 =αc2 , 22 21 22 12 22  (cid:12) 1−γ−δ(α+γ) c212− 1−α−β(α+γ) c221 = β(1−γ)−δ(1−α) c111,  G(Γ)=µ(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) (cid:0)(cid:0)11−−γα−−δβ((αα++γγ))(cid:1)(cid:1)22cc212111cc2222(cid:0)22==((δβcc111111−−cc222112))(cid:0)DD(cid:1)((ΓΓ))cc222112(cid:0)++((((βα−−11))((γδ−−11))−−α(cid:1)βδγ))cc111111(cid:1),, . (cid:12) (cid:0)1−α−β(α+γ)(cid:1) 1−γ−δ(α+γ) c2(cid:0)c2 =(βc1 −c2 ) D(Γ)c2 +((β−(cid:1)1)(γ−1)−αδ)c1 ,  (cid:12)(cid:12)(cid:12)(cid:12)(cid:12) (cid:0)(cid:0)1−α−β(α+γ)(cid:1)(cid:1)(cid:0)(cid:0)1−γ−δ(α+γ)(cid:1)(cid:1)c12111c22222 =(δc11111−c22112)(cid:0)(cid:0)D(Γ)c21221+((α−1)(δ−1)−βγ)c11111(cid:1)(cid:1)  HereandfurtherinadefinitionofasubsetofA wealwaysassumebydefaultthatµ∈A andck (i,j,k ∈{1,2})arestructure 2 2 ij constantsof µ. The followinglemma will allow us to use G(Γ) as a separatingset for some non-degenerations. Its proofis a directcalculationandsoitislefttoareader. Lemma12. ThesetG(Γ)isclosedupperinvariantandcontainsE (Γ)foranyΓ∈k4. 1 5. DEGENERATIONSOF2-DIMENSIONALALGEBRAS In thissection we describeall degenerationsof 2-dimensionalalgebras. Note that the resultsare valid for algebrasoveran algebraicallyclosedfieldofarbitrarycharacteristic. Theorem13. A hasthefollowinggraphofprimarydegenerations: 2 10 0 A4(α) B1(γ) C(α,β) D1(α,β) D3(β,γ) E1(Γ) E2(α,β,γ) E3(α,δ,ǫ) E4 γ∈{α,1−α+β} (β,γ)∈C(Γ) (β,γ)∈C(α,δ,ǫ) β+γ=0 γ= 1−δ−(1−α)ǫ 1−ǫ 1 A1(α) A2 B2(γ) D2(β,γ) 2 A3 B3 E5(α) 4 k2 Proof. Allprimarydegenerationsthatdon’tfollowfromLemma11arepresentedinthefollowingtable: degenerations parametrizedbases A (α)→A Et =te Et =t2e 1 3 1 1 2 2 A (α)→E (α) Et =e Et =e +t−1e 1 5 1 1 2 1 2 A →A Et =te Et =t2e 2 3 1 1 2 2 A →B Et =e Et =t−1e 2 3 1 1 2 2 A (α)→A Et =te −e Et =t2e 4 2 1 1 2 2 2 B (γ)→A Et =e +te Et =−t2e 1 2 1 1 2 2 2 B (γ)→B (γ) Et =te Et =e 1 2 1 1 2 2 B (γ)→A Et =e +te Et =te 2 3 1 1 2 2 1 C(α,β)→A (α) Et =te +e Et =t2e 1 1 1 2 2 2 D (α,β)→B (α) Et =te Et =e 1 2 1 1 2 2 D (α,β)→B (1−α+β) Et =te Et =e −e 1 2 1 2 2 1 2 D (α,β)→D (β,−β) Et =e Et =te 1 2 1 1 2 2 D (β,γ)→A Et =te +te Et =t2e +(β+γ)t2e 2 3 1 1 2 2 1 2 D (β,γ)→A Et = t e −e Et =te 3 2 1 1−β−γ 1 2 2 2 D (β,γ)→D (β,γ) Et =e Et =te 3 2 1 1 2 2 E (α,β,γ)→A (α) Et =te +e Et =(1−β−γ)t2e 2 1 1 1 2 2 1 E (α,δ,ǫ)→B 1−δ−(1−α)ǫ Et =te Et = ǫe1−e2 3 2 1−ǫ 1 1 2 ǫ−1 E →B (cid:16) (cid:17) Et =e −e Et =te 4 3 1 1 2 2 2 E →E (α) Et =αe +(1−α)e Et =(α−t)e +(1−α+t)e 4 5 1 1 2 2 1 2 Thenexttabledescribesseparatingsetsforallrequirednon-degenerationsand,thus,finishestheproofofthetheorem.

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.