Автор: Hertling С.  

Теги: mathematics  

ISBN: 978-0-511-54310-4

Год: 2002

Текст
                    Frobenius manifolds and moduli
spaces for singularities
CLAUS HERTLING
Cambridge
UNIVERSITY PRESS


CAMBRIDGE TRACTS IN MATHEMATICS General Editors B. BOLLOBAS, W. FULTON, А. КАТОК, F. KIRWAN, P. SARNAK 151 Frobenius manifolds and moduli spaces for singularities
Contents Preface . page viii Part 1. Multiplication on the tangent bundle 1 Introduction to part 1 3 1.1 First examples 4 1.2 Fast track through the results 5 2 Definition and first properties of F-manifolds 9 2.1 Finite-dimensional algebras 9 2.2 Vector bundles with multiplication 11 2.3 Definition of F-manifolds 14 2.4 Decomposition of F-manifolds and examples 16 2.5 F-manifolds and potentiality 19 3 Massive F-manifolds and Lagrange maps 23 3.1 Lagrange property of massive F-manifolds 23 3.2 Existence of Euler fields 26 3.3 Lyashko-Looijenga maps and graphs of Lagrange maps 29 3.4 Miniversal Lagrange maps and F-manifolds 32 3.5 Lyashko-Looijenga map of an F-manifold 35 4 Discriminants and modality of F-manifolds 40 4.1 Discriminant of an F-manifold 40 4.2 2-dimensional F-manifolds 44 4.3 Logarithmic vector fields 47 ix 4.4 Isomorphisms and modality of germs of F-manifolds 52 \\ 4.5 Analytic spectrum embedded differently 56
vi Contents 5 Singularities and Coxeter groups 61 5.1 Hypersurface singularities 61 5.2 Boundary singularities 69 5.3 Coxeter groups and F-manifolds 75 5.4 Coxeter groups and Frobenius manifolds 82 5.5 3-dimensional and other F-manifolds 87 Part 2. Frobenius manifolds, GauB-Manin connections, and moduli spaces for hypersurface singularities 6 Introduction to part 2 99 6.1 Construction of Frobenius manifolds for singularities 100 6.2 Moduli spaces and other applications 104 7 Connections over the punctured plane 109 7.1 Flat vector bundles on the punctured plane 109 7.2 Lattices 113 7.3 Saturated lattices 116 7.4 Riemann-Hilbert-Birkhoff problem 120 7.5 Spectral numbers globally 128 8 Meromorphic connections 131 8.1 Logarithmic vector fields and differential forms 131 8.2 Logarithmic pole along a smooth divisor 134 8.3 Logarithmic pole along any divisor 139 8.4 Remarks on regular singular connections 143 9 Frobenius manifolds and second structure connections 145 9.1 Definition of Frobenius manifolds 145 9.2 Second structure connections 148 9.3 First structure connections 154 9.4 From the structure connections to metric and multiplication 157 9.5 Massive Frobenius manifolds 160 10 GauB-Manin connections for hypersurface singularities 165 10.1 Semiuniversal unfoldings and F-manifolds 165 10.2 Cohomology bundle 167 10.3 GauB-Manin connection 170 10.4 Higher residue pairings 179 10.5 Polarized mixed Hodge structures and opposite filtrations 183 10.6 Brieskorn lattice 188 Contents 11 Frobenius manifolds for hypersurface singularities 11.1 Construction of Frobenius manifolds 11.2 Deformed flat coordinates 11.3 Remarks on mirror symmetry 11.4 Remarks on oscillating integrals 12 д-constant stratum 12.1 Canonical complex structure 12.2 Period map and infinitesimal Torelli 13 Moduli spaces for singularities 13.1 Compatibilities 13.2 Symmetries of singularities 13.3 Global moduli spaces for singularities 14 Variance of the spectral numbers 14.1 Socle field 14.2 G-function of a massive Frobenius manifold 14.3 Variance of the spectrum Bibliography Index vn 195 195 205 211 212 218 218 224 230 230 235 240 248 248 251 256 260 269
Preface Frobenius manifolds are complex manifolds with a rich structure on the holo- morphic tangent bundle, a multiplication and a metric which harmonize in the most natural way. They were defined by Dubrovin in 1991, motivated by the work of Witten, Dijkgraaf, E. Verlinde, and H. Verlinde on topological field theory. Originally coming from physics, Frobenius manifolds now turn up in very different areas of mathematics, giving unexpected relations between them, in quantum cohomology, singularity theory, integrable systems, symplectic ge- geometry, and others. The isomorphy of certain Frobenius manifolds in quantum cohomology and in singularity theory is one version of mirror symmetry. This book is devoted to the relations between Frobenius manifolds and sin- singularity theory. It consists of two parts. In part 1 F-manifolds are studied, manifolds with a multiplication on the tangent bundle with a natural integrability condition. They were introduced in [HM][Man2, I§5]. Frobenius manifolds are F-manifolds. Studying F-manifolds, one is led directly to discriminants, a classical subject of singularity theory, and to Lagrange maps and their singularities. Our development of the general struc- structure of F-manifolds is at the same time an introduction to discriminants and Lagrange maps. As an application, we use some work of Givental to prove a conjecture of Dubrovin about Frobenius manifolds and Coxeter groups. In part 2 we take up the construction of Frobenius manifolds in singularity theory. Already in 1983 K. Saito and M. S aito had found that the base space of a semiuniversal unfolding of an isolated hypersurface singularity can be equipped with the structure of a Frobenius manifold. Their construction involves the GauB-Manin connection, polarized mixed Hodge structures, K. Saito's higher residue pairings, and his primitive forms. It was hardly accessible for nonspe- cialists. We give a more elementary detailed account of the construction, explain all ingredients, and develop or cite all necessary results. via Preface IX We give a number of applications. The deepest one is the construction of global moduli spaces for isolated hypersurface singularities. The construction of K. Saito and M. Saito is related to a recent construction of Frobenius manifolds via oscillating integrals by Sabbah and Barannikov. We comment upon that. Background and other books. The reader should know the basic concepts of complex analytic geometry, including coherent sheaves and flatness (cf. for example [Fi]). All notions from symplectic geometry which are used can be found in [AGV1, chapter 18]. An excellent basic reference on flat connections and vector bundles (and much more) is the forthcoming book [Sab4]. It also gives a viewpoint on Frobenius manifolds which complements ours. Two fun- fundamental books on Frobenius manifolds are [Du3] and [Man2]. Our treatment of singularities and their GauB-Manin connection is essentially self-contained and gives precisely what is needed, but it is quite compact. Some books which expound several aspects in much more detail are [AGVl][AGV2][Ku][Lo2]. Acknowledgements. This book grew out of my habilitation. I would like to thank many people. E. Brieskorn was my teacher in singularity theory and defined in 1970 the wonderful object Щ', which is now called the Brieskorn lattice. Yu. Manin introduced me to Frobenius manifolds. The common paper [HM] was the starting point for part 1. His papers and those of B. Dubrovin and C. Sabbah, and discussions with them were very fruitful. G.-M. Greuel and G. Pfister sharpened my view of moduli problems. M. Schulze and M. Rosellen made useful comments. Of course, this book builds on the work of many people in singularity the- theory; Arnold, Givental, Looijenga, Malgrange, K. Saito, M. Saito, Scherk, O.P. Shcherbak, Slodowy, Steenbrink, Teissier, Varchenko, Wall, and many others. A good part of the book was written during a stay at the mathematics depart- department of the University Paul Sabatier in Toulouse. I thank the department and especially J.-F. Mattel for their hospitality. Bonn, July 2001 Claus Hertling
Parti Multiplication on the tangent bundle
Chapter 1 Introduction to part 1 An F-manifold is a complex manifold M such that each holomorphic tangent space TtM,t e M, is a commutative and associative algebra with unit element, and the multiplication varies in a specific way with the point t e M. More precisely, it is a triple (M, о, ё) where о is an O^-bilinear commutative and associative multiplication on the holomorphic tangent sheaf Тм, е is a global unit field, and the multiplication satisfies the integrability condition LieXoy(o) = X о Liey(o) + Y о Liex(o) A.1) for any two local vector fields X and Y in Тм. This notion was first defined in [HM][Man2, I§5], motivated by Frobenius manifolds. Frobenius manifolds are F-manifolds. Part 1 of this book is devoted to the local structure of F-manifolds. It turns out to be closely related to singularity theory and symplectic geometry. Discri- Discriminants and Lagrange maps play a key role. In the short section 1.1 of this introduction the reader can experience some of the geometry of F-manifolds. We sketch a construction of 2-dimensional F-manifolds which shows how F-manifolds turn up 'in nature' and how they are related to discriminants. In section 1.2 we offer a fast track through the main notions and results of chapters 2 to 5. In chapters 2 to 4 the general structure of F-manifolds is developed. In chapter 5 the most important classes of F-manifolds are discussed. In chapter 2 F-manifolds are defined and some basic properties are estab- established. One property shows that F-manifolds decompose locally in a nice way. Another one describes the relation to connections, metrics, and the potentiality condition of Frobenius manifolds. In chapter 3 the relation to symplectic geometry and especially to Lagrange maps is discussed. This allows use to be made of Givental's paper [Gi2] on singular Lagrange varieties and their Lagrange maps.
4 Introduction to part I Chapter 4 presents several notions and results, which are mostly motivated by corresponding notions and results in singularity theory. Most important are the discriminants and their geometry. In chapter 5 F-manifolds from hypersurface singularities, boundary singu- singularities, and Coxeter groups are discussed. In the case of Coxeter groups we extend some results of Givental [Gi2] and use them to prove a conjecture of Dubrovin about their Frobenius manifolds. The reader should have the following background. There should be fa- familiarity with the basic concepts of complex analytic geometry, including coherent sheaves and flatness. One reference is [Fi]. There should also be awareness of those notions from symplectic geometry which are treated in [AGV1, chapter 18] (canonical 1-form on the cotangent bundle, Lagrange fibration, Lagrange map, generating function). We recommend this reference. In chapter 5 some acquaintance with singularity theory makes the reading easier, but it is not necessary. Good references are [AGV1] and [Lo2]. 1.1 First examples To give the reader an idea of what F-manifolds look like and how they arise naturally, a construction of 2-dimensional F-manifolds is sketched. A systematic treatment is given in sections 4.1 and 4.2. Let W be a finite Coxeter group of type him), m > 2, acting on R2 and (by C-linear extension) on C2. Then the ring С [jci,jc2]1v С C[x\,X2]oiW -invariant polynomials is C[jc] ,x{\w = C[fi, ti\ with 2 homogeneous generators t\ and t2 of degrees m and 2. Therefore the quotient space C2/ W =: M is isomorphic to C2 as an affine algebraic variety, and the vector field e := ^- is unique up to multiplication by a constant. The image in M of the union of the complexified reflection hyperplanes is the discriminant V. We choose t\ and t-i such that it is given as V = {t e M \ t\ - ^ff = 0}. For a point t e M with Гг ф 0, the pair (e, V) gives rise to a multiplication on T,M in the following way, which is illustrated in figure 1.1. The e-orbit through the point t intersects the discriminant at 2 points. We shift the tangent hyperplanes of V at these points with the flow of e to TtM. We find that they are transverse to one another and to e. Therefore there are 2 unique vectors e\ and e% in T,M which are tangent to these lines and satisfy е = е\+в2- We define a multiplication on T, M by e,- о ej := 5y e,-. It is obviously commutative and associative, and e is the unit vector. If we write this multiplication in terms of the coordinate fields e :— ?- and ^-, after some calculation we find 4г о 4г = *о~2 ¦ irr> and e is the Ot2 0/2 O?2 ^ 3/i ' 1.2 Fast track through the results Figure 1.1 unit field. Therefore the multiplication extends holomorphically to the whole tangent bundle TM. One can show that it satisfies A.1). The orbit space M is an F-manifold. This construction of an F-manifold from a discriminant V and a transver- transversal vector field e extends to higher dimensions (Corollary 4.6) and yields F-manifolds in many other cases, for example for all finite Coxeter groups (section 5.3). 1.2 Fast track through the results The most notable (germs of) F-manifolds with many typical and some special properties are the base spaces of semiuniversal unfoldings of isolated hyper- hypersurface singularities and of boundary singularities (sections 5.1 and 5.2). Here the tangent space at each parameter is canonically isomorphic to the sum of the Jacobi algebras of the singularities above this parameter. Many of the general results on F-manifolds have been known in another guise in the hypersurface singularity case and all should be compared with it. One reason why the integrability condition A.1) is natural is the following: Let (M, p) be the germ of an F-manifold (M, o, e). The algebra TPM decom- decomposes uniquely into a sum of (irreducible) local algebras which annihilate one another (Lemma 2.1). Now the integrability condition A.1) ensures that this infinitesimal decomposition extends to a unique decomposition of the germ (M, p) into a product of germs of F-manifolds (Theorem 2.11). If the multiplication at TpM is semisimple, that is, if TpM decomposes into 1 -dimensional algebras, then this provides canonical coordinates и \,..., un on (M, p) with g|: о g2_ = 5^. _i-. in fact, at points with semisimple multiplication the integrability condition A.1) is equivalent to the existence of such canonical coordinates. In the hypersurface case, the decomposition of the germ (M, p)
6 Introduction to part 1 for some parameter p is the unique decomposition into a product of base spaces of semiuniversal unfoldings of the singularities above p. Another reason why A.1) is natural is its relation to the potentiality of Frobenius manifolds. There exist F-manifolds such that not all tangent spaces are Frobenius algebras. They cannot be Frobenius manifolds. But if all tangent spaces are Frobenius algebras then the integrability condition A.1) is related to a version of potentiality which requires a metric on M that is multiplication invariant, but not necessarily flat. See section 2.5 for details. The most important geometric object which is attributed to an n-dimensional manifold M with multiplication о on the tangent sheaf Тм and unit field e (with or without A.1)) is the analytic spectrum L := Specan(T^) с T*M (see section 2.2). The projection ж : L -> M is flat and finite of degree n. The fibre n~\p) С L above p e M consists of the set Homaig(TpM, C) of algebra homomorphisms from Tp M to C; they correspond 1 -1 to the irreducible subalgebras of (TpM, o) (see Lemma 2.1). The multiplication on TM can be recovered from L, because the map a : T M X н> a(X)\L A.2) is an isomorphism of CV-algebras; here X is any lift of X to T*M and a is the canonical 1-form on T*M. The values of the function a(X) on л~1(р) are the eigenvalues of Xo : TPM -» TpM. The analytic spectrum Lisa reduced variety if and only if the multiplication is generically semisimple. Then the manifold with multiplication (M, o, e) is called massive. Now, a third reason why the integrability condition A.1) is natural is this: Suppose that (M, o, e) is a manifold with generically semisimple multiplication. Then L с T*M is a Lagrange variety if and only if (M, o, <?) is a massive F-manifold (Theorem 3.2). The main body of part 1 is devoted to the study of germs of massive F-manifolds at points where the multiplication is not semisimple. We will make use of the theory of singular Lagrange varieties and their Lagrange maps, which has been worked out by Givental in [Gi2]. In fact, the notion of an irreducible germ (with respect to the above decomposition) of a massive F-manifold is equivalent to Givental's notion of a miniversal germ of a flat Lagrange map (Theorem 3.16). Via this equivalence Givental's paper contains many results on massive F-manifolds and will be extremely useful. Locally the canonical 1-form a on T*M can be integrated on the ana- analytic spectrum L of a massive F-manifold (M, o, e) to a generating function F : L -? С which is continuous on L and holomorphic on Lreg. It depends on a property of L, which is weaker than normality or maximality of the complex structure of L, whether F is holomorphic on L (see section 3.2). 1.2 Fast track through the results 7 If F is holomorphic on L then it corresponds via A.2) to an Euler field E — a~\F) of weight 1, that is, a vector field on M with Lie?(o) = о (Theorem 3.3). In any case, a generating function F : L -> С gives rise to a Lyashko- Looijenga map Л : M -» C" (see sections 3.3 and 3.5) and a discriminant V=tt(F-](O))cM. If F is holomorphic and an Euler field E = a~](F) exists then the discrimi- discriminant V is the hypersurface of points where the multiplication with E is not in- vertible.ThenitisafreedivisorwithlogarithmicfieldsDer/M(log'P) = EoTM (Theorem 4.9). This generalizes a result of K. Saito for the hypersurface case. From the unit field e and a discriminant V С M one can reconstruct every- everything. One can read off the multiplication on TM in a very nice elementary way (Corollary 4.6 and section 1.1): The e-orbit of a generic point p e M intersects V at n points. One shifts the n tangent hyperplanes with the flow of e to TpM. Then there exist unique vectors e\(p),..., en(p) e TPM such that ?"=i ei(P) = e(p) and YH=\ c • е<(^) = tpm and such Ла1 the subspaces ?\ ,. С • e,{p), j = 1,..., n, are the shifted hyperplanes. The multiplication on TPM is given by <?,(/?) о ej(p) = 3,7e,-(p). In the case of hypersurface singularities and boundary singularities, the clas- classical discriminant in the base space of a semiuniversal unfolding is such a discriminant. The critical set С in the total space of the unfolding is canonically isomorphic to the analytic spectrum L; this isomorphism identifies the map a in A.2) with a Kodaira-Spencer map ac : TM -> (пс)*Ос and a generating function F : L —> С with the restriction of the unfolding function to the critical set C. This Kodaira-Spencer map ac is the source of the multiplication on TM in the hypersurface singularity case. The multiplication on TM had first been defined in this way by K. Saito. Critical set and analytic spectrum are smooth in the hypersurface singular- singularity case. By the work of Arnold and Hormander on Lagrange maps and sin- singularities an excellent correspondence holds (Theorem 5.6): each irreducible germ of a massive F-manifold with smooth analytic spectrum comes from an isolated hypersurface singularity, and this singularity is unique up to stable right equivalence. By the work of Nguyen huu Due and Nguyen tien Dai the same corre- correspondence holds for boundary singularities and irreducible germs of massive F-manifolds whose analytic spectrum is isomorphic to (C1, 0) x ({(*, у) б C2 | xy = 0}, 0) with ordered components (Theorem 5.14). The complex orbit space M := C/ W = C" of a finite irreducible Coxeter group W carries an (up to some rescaling) canonical structure of a massive F-manifold: A generating system P\,...,Pn of W-invariant homogeneous
8 Introduction to part 1 polynomials induces coordinates (| („onM. Precisely one polynomial, e.g. Pi, has highest degree. The field щ is up to a scalar independent of any choices. This field e := ^ as the unit field and the classical discriminant V С M, the image of the reflection hyperplanes, determine in the elementary way described above the structure of a massive F-manifold. This follows from [Du2][Du3, Lecture 4] as well as from [Gi2, Theorem 14]. Dubrovin established the structure of a Frobenius manifold on the complex orbit space M = C/ W, with this multiplication, with K. Saito's flat metric on M, and with a canonical Euler field with positive weights (see Theorem 5.23). At the same place he conjectured that these Frobenius manifolds and their products are (up to some well-understood rescalings) the only massive Frobenius man- manifolds with an Euler field with positive weights. We will prove this conjecture (Theorem 5.25). Crucial for the proof is Givental's result [Gi2, Theorem 14]. It characterizes the germs (M, 0) of F-manifolds of irreducible Coxeter groups by geometric properties (see Theorem 5.21). We obtain from it the following intermediate result (Theorem 5.20): An irreducible germ (M, p) of a simple F-manifold such that Tp M is a Frobenius algebra is isomorphic to the germ at 0 of the F-manifold of an irreducible Coxeter group. A massive F-manifold (M, o, e) is called simple if the germs (M, p), p € M, of F-manifolds are contained in finitely many isomorphism classes. Theorem 5.20 complements in a nice way the relation of irreducible Coxeter groups to the simple hypersurface singularities An, Dn, En and the simple boundary singularities Bn,Cn,F4. In dimensions 1 and 2, up to isomorphism all the irreducible germs of massive F-manifolds come from the irreducible Coxeter groups Ai and I2(m) (m > 3) with /2C) = A2, /2D) = B2, /2E) =: Я2, /2F) = G2. But already in dime- dimension 3 the classification is vast (see section 5.5). Chapter 2 Definition and first properties of F-manifolds An F-manifold is a manifold with a multiplication on the tangent bundle which satisfies a certain integrability condition. It is defined in section 2.3. Sections 2.4 and 2.5 give two reasons why this is a good notion. In section 2.4 it is shown that germs of F-manifolds decompose in a nice way. In section 2.5 the relation to connections and metrics is discussed. It turns out that the integrability condition is part of the potentiality condition for Frobenius manifolds. Therefore Frobenius manifolds are F-manifolds. Section 2.1 is a self-contained elementary account of the structure of finite dimensional algebras in general (e.g. the tangent spaces of an F-manifold) and Frobenius algebras in particular. Section 2.2 discusses vector bundles with multiplication. There the caustic and the analytic spectrum are defined, two notions which are important for F-manifolds. 2.1 Finite-dimensional algebras In this section (Q, o, e) is a C-algebra of finite dimension (> 1) with commuta- commutative and associative multiplication and with unit e. The next lemma gives precise information on the decomposition of Q into irreducible algebras. The statements are well known and elementary. They can be deduced directly in the given order or from more general results in commutative algebra (Q is an Artin algebra). Algebra homomorphisms are always supposed to map the unit to the unit. Lemma 2.1 Let (Q, o, e) be as above. As the endomorphisms xo : Q -» Q, x € Q, commute, there is a unique simultaneous decomposition Q = ф^=1 Qk into generalized eigenspaces Qk (with dimc Qk > U Define ek e Qk by e = El=i ek- Then (i) One has Qj о Qk = Ofor j ф к; also ek ф 0 and ej о ек = 8]кек; the element ek is the unit of the algebra Qk = ek о Q.
10 Definition and first properties of F-manifolds (ii) The function Xk : Q -» С which associates to x € Q the eigenvalue of xo on Qk is an algebra homomorphism; Xj ф Xkforj ф к. (iii) The algebra (Qk, o, ek) is an irreducible and a local algebra with maximal ideal mk = Qk П ker(A.*). (iv) The subsets ker(Xt) = m^ ф 0^ Qj,k = l,..,l, are the maximal ideals of the algebra Q; the complement Q — \Jk ker(A.*) is the group ofinvertible elements of Q. (v) Theset{Xu...,Xi}^ Homc_a^(<2, C). (vi) The localization <2кег(щ is isomorphic to Qk. We call this decomposition the eigenspace decomposition of (Q,o,e). The set L := {X\,..., А./} С Q* carries a natural complex structure Ol such that OL{L) = Q and 0/.,^ = ?>*. More details on this will be given in section 2.2. The algebra (or its multiplication) is called semisimple if Q decomposes into 1-dimensional subspaces, Q = 0^ Qk = ®f™te С • ek. An irreducible algebra Q = С • e ф m with maximal ideal m is a Gorenstein ring if the socle AnnG(m) has dimension 1. An algebra Q = 0^=1 Qk is a Frobenius algebra if each irreducible sub- algebra is a Gorenstein ring (cf. for example [Kun]). The next (also well known) lemma gives equivalent conditions and additional information. Note that this classical definition of a Frobenius algebra is slightly weaker than Dubrovin's: he calls an algebra (Q, o, e) together with a. fixed bilinear form g as in Lemma 2.2 (a) (iii) a Frobenius algebra. Lemma 2.2 (a) The following conditions are equivalent. (i) The algebra (Q, o, e) is a Frobenius algebra, (ii) As a Q-module Hom(<2, C) = Q. (iii) There exists a bilinear form g : Q x Q -> С which is symmetric, non- degenerate and multiplication invariant, i.e. g(a о b, c) = g(a, b о с). (b) Let Q = 0j=| Qk be a Frobenius algebra and Qk = С • ek ф тк. The generators of Hom(?>, C) as a Q-module are the linear forms f:Q-*-C with /(AnnGt (щО) = С for all k. One obtains a 1-1 correspondence between these linear forms and the bi- bilinear forms g as in (a) (iii) by putting g(x, y) := f(x о у). Proof: (a) Any of the conditions (i), (ii), (iii) in (a) is satisfied for Q if and only if it is satisfied for each irreducible subalgebra Qk. One checks this with Qj ° Qk = 0 for j' ф к. So we may suppose that Q is irreducible. 2.2 Vector bundles with multiplication 11 (i) <;—>¦ (ii) A linear form /eHomB,C) generates HomB,C) as a g-module if and only if the linear form (x i->- f(y о х)) is nontrivial for any у € Q - {0}, that is, if and only if f(y о Q) - С for any у e Q - {0}. The socle AnnQ(m) is the set of the common eigenvectors of all endomor- phismsxo : Q -> Q, x € Q.Ifdim Ann2(m) > 2 then for any linear form / an element у 6 (ker / П Ann2(m)) - {0} satisfies у о Q — С • у and f(y о Q) = 0; so / does not generate Hom(<2, C). If dim Ann2(m) = 1 then it is contained in any nontrivial ideal, because any such ideal contains a common eigenvector of all endomorphisms. The set у о Q for у € Q - {0} is an ideal. So, then a linear form / with /(Апп2(т)) = С generates HomB, C) as a Q-module. (i) =$¦ (iii) Choose any linear form / with /(AnnG(m)) = С and define g by g(x, y) ."= f(x о у). It remains to show that g is nondegenerate. But for any x € Q - {0} there exists а у € Q with С • x о у = AnnG(m), because AnnG(m) is contained in the ideal xo Q. (iii) =»(i) The equalities g(m, Ann2(m)) = g(e,moAnn2(m)) = g(e, 0) = 0 imply dim Anne(m) = 1. (b) Starting with a bilinear form g, the corresponding linear form / is given by f(x) = g(x, e). The rest is clear from the preceding discussion. Q The semisimple algebra Q = (Bf™® С • e^is a Frobenius algebra. A classical result is that the complete intersections Сс™,о/(/ь • • •. fm) are Gorenstein. But there are other Gorenstein algebras, e.g. C{x, y, z}/(x2, y2,xz, yz, xy — z2) is Gorenstein, but not a complete intersection. Finally, in the next section vector bundles with multiplication will be consid- considered. Condition (ii) of Lemma 2.2 (a) shows that there the points whose fibres are Frobenius algebras form an open set in the base. 2.2 Vector bundles with multiplication Now we consider a holomorphic vector bundle Q —> M on a complex manifold M with multiplication on the fibres: The sheaf Q of holomorphic sections of the bundle Q -> M is equipped with an См-bilinear commutative and associative multiplication о and with a global unit section e. The set (JpeM HomC-aig(Q(p), C) of algebra homomorphisms from the sin- single fibres Q(p) to С (which map the unit to 1 6 C) is a subset of the dual bundle Q* and has a natural complex structure. It is the analytic spectrum Specan(Q). We sketch the definition ([Hou, ch. 3], also [Fi, 1.14]): The См-sheaf SymOM Q can be identified with the Cw-sheaf of holomorphic functions on Q* which are polynomial in the fibres. The canonical CM-algebra homomorphism SymOM Q -*¦ Q which maps the multiplication in SymOM Q
12 Definition and first properties of F-manifolds to the multiplication о in Q is surjective. The kernel generates an ideal I in OQ.. One can describe the ideal locally explicitly: suppose U С М is open and &\,..., &„ e Q([/)is abase of sections of the restriction Q | v -> U with 5, = e and S; о Sj = ?t af,<5b denote by y\,..., yn the fibrewise linear functions on Q*\U which are induced by <5i,..., <5n; then the ideal I is generated in Q*\U by -1 and B.1) The support of Oq.IT with the restriction of OQ./2 as structure sheaf is Specan(Q) С Q*. We denote the natural projections by jzq. :Q*-+M and 7Г : Specan(Q) -*¦ M. A part of the following theorem is already clear from the discussion. A complete proof and thorough discussion can be found in [Hou, ch. 3]. Theorem 2.3 The set \JpeM Homc_afe (?)(», C) is the support of the analytic spectrum Specan(Q) =: L. The composition of maps a : Q *-+ {tzq.)*Oq. -> 7t»OL B.2) is an isomorphism of О M-algebras and of free О м -modules of rank n, here n is the fibre dimension ofQ-*M. The projection n : L -*¦ M is finite and flat of degree n. Consider a point p e M with eigenspace decomposition Q(p)= ®l=i Qk(p) and L П 7r~' (p) = {Х\,..., ХцР)}. The restriction of the isomorphism Up) to the fibre over p yields isomorphisms Qk(p) = OL С B.3) B.4) Corollary 2.4 In a sufficiently small neighbourhood U of a point p e M, the eigenspace decomposition Q(p)= ф'^, Qk(p) of the fibre Q(p) extends uniquely to a decomposition of the bundle Q\y —> U into multiplication in- invariant holomorphic subbundles. Proof of Corollary 2.4: The OMp-free submodules OLikk in the decomposition in B.3) of (я*Оь)р are obviously multiplication invariant. Via the isomorphism a one obtains locally a decomposition of the sheaf Q of sections of Q -*¦ M into multiplication invariant free CV-submodules. D 2.2 Vector bundles with multiplication 13 Of course, the induced decomposition of Q (q) for a point q in the neighbour- neighbourhood of p may be coarser than the eigenspace decomposition of Q(q). The base is naturally stratified with respect to the numbers and dimensions of the components of the eigenspace decompositions of the fibres of Q ->¦ M. To make this precise we introduce a partial ordering >- on the set V of partitions of n: !«« p = (Pi,..., Pm) | Pi e N, Pi > pi+i, J^Pi = for P, у е V define 3or:{l,...,/(y)}-Ml,...,/(/*)} s.t Pj = J^ y,. The partition P(p) of a fibre Q(p) is the partition of n = dim Q(p) by the dimensions of the subspaces of the eigenspace decomposition. Proposition 2.5 Fix a partition P e V. The subset {p e M | P{p) >- P\ is empty or an analytic subset of M. Proof. The partition P{f) of a polynomial of degree n is the partition of n by the multiplicity of the zeros of /. Fact: The space {a e C" | P(zn + ?"=i щг"~') >¦ P) is an algebraic subvariety of С with normalization isomorphic to C'w. For the proof one only has to regard the finite map C->-C\ иь> ((-1)'<Т|(и))|=1,..,п (сГ|(") is the i-th symmetric polynomial). A section X &Q(U),V CM open, induces via the coefficients of the charac- characteristic polynomial pch,xo of multiplication by X a holomorphic map U -*¦ C. Hence the set {q e U | P(pch,Xo) > P\ is analytic. The intersection of such analytic sets for a basis of sections in U is {q e [/ | P(q) >- P). ? Suppose that M is connected. Then there is a unique partition Po such that (p 6 M | P(p) = j8o) is open in M. The complement К := [p 6 M \ P{p) ф Po\ will be called the caustic in M; this name is motivated by the Lagrange maps (sections 3.1, 3.3, 3.4) and the hypersurface singularities (section 5.1). The multiplication is generically semisimple if and only if/Jo = (l,...,l). Proposition 2.6 The caustic K, is a hypersurface or empty. Proof. Locally in M — K, there is a holomorphically varying decomposition Q(p) = ф^=1 Qk(p) with partial unit fields e\,..., e/. Suppose dim(X], p) < dim M -2 for some point p e K.. Then in a neighbour- neighbourhood U the complement U — K. is simply connected. There is no monodromy
14 Definition and first properties of F-manifolds for the locally defined vector fields e\,..., et in U - K.. They extend to vector fields in U. For p e U — K. the map e^o : Q(p)-+ Q{p) is the projec- projection to Qk(p)- Because of e\ + ¦ ¦ ¦ + <?/ = e these projections extend to all of U and yield a decomposition of Q(p) as above for all p € U. Hence /С П U = 0. ? 2.3 Definition of F-manifolds An F-manifold is a manifold Л/ with a multiplication on the tangent bundle TM which harmonizes with the Lie bracket in the most natural way. A neat formulation of this property requires the Lie derivative of tensors. Remark 2.7 Here the sheaf of (k, /)-tensors (k, I e No) on a manifold M is the sheaf of C>A/-module homomorphisms HomoM@*=1 TM, ®|=1 Тм). А @, /)-tensor T : OM -*¦ ®|=, TM can be identified with 7"A). Vector fields are @, l)-tensors, a (commutative) multiplication on TM is a (symmetric) B,1)- tensor. The Lie derivative Liex with respect to a vector field X is a derivation on the sheaf of (k, Z)-tensors. It is Liex(/) = X(f) for functions /, Uex(Y) = [X, Y] for vector fields Y, Uex(Yi ® ... ® Я) = ?,- F] <g> ..[X, F;].. <g> Y, for @, /)- tensors, and (Liex7")(F) = Liex(T(Y))- T(Liex(Y))for(k, /)-tensors T. One can always write it explicitly with Lie brackets. Because of the Jacobi identity the Lie derivative satisfies Lie^y] == [Liex, Liey]. Definition 2.8 (a) An F-manifold is a triple (M, o, e) where M is a com- complex connected manifold of dimension > 1, о is a commutative and associative Ом-bilinear multiplication Тм х Тм -*¦ Тм, е is a global unit field, and the multiplication satisfies for any two local vector fields X, Y Liew(o) = X о Liey(o) + Y о Liex(o). B.5) (b) Let (M, o, e) be an F-manifold. An Euler field E of weight d € С is a global vector field E which satisfies Lie?(o) = d ¦ o. B.6) (If no weight is mentioned, an Euler field will usually mean an Euler field of weight 1.) Remarks 2.9 (i) We do not require that the algebras (TpM,o,e(p)) are Frobenius algebras (cf. section 2.1). Nevertheless, this is a distinguished class. Frobenius manifolds are F-manifolds [HM][Manl, I§5]. This will be discussed in section 2.5. 2.3 Definition of F-manifolds 15 (ii) Definition 2.8 differs slightly from the definition of F-manifolds in [HM] by the addition of a global unit field e. This unit field is important, for example, for the definition of SpecanG^/). Also, the Euler fields were called weak Euler fields in [HM] in order to separate them from the Euler fields with stronger properties of Frobenius manifolds. This is not necessary here. (iii) Formula B.5) is equivalent to [X о Y, Z о W] - [X о Y, Z] о W - [X о Y, W] о Z - Y о [X, Z о W] + Y о [X, Z] о W + Y о [X, W] о Z = 0 B.7) for any four (local) vector fields X, Y, Z, W. Formula B.6) is equivalent to [E, X о Y] - [E, X] о Y - X о [E, Y] - d ¦ X о Y = 0 B.8) for any two (local) vector fields X, Y. The left hand side of B.8) is OM- polylinear with respect to X and Y, because Liee(o) is a B, l)-tensor. The left hand side of B.7) is Ом-polylinear with respect to X, Y, Z, W. Hence it defines a D, l)-tensor. In order to check B.5) and B.6) for a manifold with multiplication, it suffices to check B.7) and B.8) for a basis of vector fields. (iv) The unit field e in an F-manifold (M, o, e) plays a distinguished role. It is automatically nowhere vanishing. It is an Euler field of weight 0, Liee(o) = 0-o, B.9) because of B.5) for X = F = e. So, the multiplication of the F-manifold is constant along the unit field. (v) An Euler field E of weight d ф 0 is not constant along the unit field. But one has for any d eC [e,E] = d-e, B.10) because of B.8) for X = F = e. More generally, in [HM][Manl, I§5] the identity [E°n, Eom] = d(m - n) ¦ B.11) is proved. Section 3.1 will show how intrinsic the notion of an Euler field is for an F-manifold. (vi) The sheaf of Euler fields of an F-manifold (M, o, e) is a Lie subalgebra of TM. If E{ and E2 are Euler fields of weight d\ and d2, then с ¦ E\ (c € C) is an Euler field of weight с • d\, E\ + Ег is an Euler field of weight d\ + d2, and [Ei, E2] is an Euler field of weight 0. The last holds because of Lie[?,?2] = [Lie?l, Lie?J (cf. Remark 2.7 and [HM][Manl, I§5]).
16 Definition and first properties of F-manifolds (vii) The caustic К of an F-manifold is the subvariety of points p e M such that the eigenspace decomposition of TpM has fewer components than for generic points (cf. section 2.2). The caustic is invariant with respect to e because of B.9). 2.4 Decomposition of F-manifolds and examples Proposition 2.10 The product oftwo F-manifolds (M\, ou e\)and(M2, o2, e2) is an F-manifold (M, o, e) = (M\ x M2, o\ ф о2, e\ + e2). If E\ and Ei are Eider fields on M\ and M2 of the same weight d then the sum E\ + E2 (of the lifts to M) is an Eider field of weight d on M. Proof. The tangent sheaf decomposes, Та = OM ¦ pr^TMl © OM ¦ pr^TMl. Any vector fields Xit У, e prJ~lTMl, i = 1, 2, satisfy X, о Y, e prr'TMi, [X,, Y,] e Pr7lTMl, XioY2 = 0, [XuY2] = 0. This together with B.7) for vector fields in Тщ and for vector fields in Тмг gives B.7) for vector fields in prj"" lTMl U pr^T^. Because of the O^-polylinearity then B.7) holds for all vector fields. For the same reasons, E\ + E2 satisfies B.8). D Theorem 2.11 Let (M, p) be the germ in p e M of an F-manifold (M, o, e). Then the eigenspace decomposition TpM — ®'k=i(TpM)k of the algebra Tp M extends to a unique decomposition 1 (M, p) = Y\(Mk, p) k=\ ofthegerm(M, p) into a product of germs of F-manifolds. These germs (Mk, p) are irreducible germs of F-manifolds, as the algebras TpMk = (TpM)k are already irreducible. An Euler field E on (M, p) decomposes into a sum of Eulerfields of the same weights on the germs (Mk, p) of F-manifolds. Proof. By Corollary 2.4 the eigenspace decomposition of TpM extends in some neighbourhood of p to a decomposition of the tangent bundle into a sum of 2.4 Decomposition of F-manifolds and examples 17 multiplication invariant subbundles. First we have to show that these subbundles and any sum of them are integrable. Accordingly, let TMtP = ®k=i(TMtP)k be the decomposition of TM<p into multiplication invariant free 0MiP-submodules, and e = J^k ek with ek е (TM,P)k- Then eko : TM<p ->¦ (TMiP)k is the projection; ey oek= 8jkek. Claim: (i) With respect to ek the multiplication is invariant, Lieet(o) = 0 • o; (ii) the vectorfields ej and ek commute, [ej, ek] = 0; (Hi) they leave the subsheaves invariant, [ej, (TMtP)k] С (ТМ:Р)к; (iv) the subsheaves satisfy [(TMiP)j, (Тм,Р)к\ С (JmiP)j + (TM,p)k- Proof of the claim: (i) The equality Sjk ¦ Lieet(o) = Lieejoet(o) = e,- о Lie?l(o) + ek о Lieej(o). implies for j ф к as well as for j = к that e,- о Liee4(o) = 0 ¦ o. Thus (ii) The equality 0 = Lieej.(o)(eb ek) — [ej, ek о ek] — 2ek о [ej, ek~\ shows that [ejt ek] e (TMtP)k, so for j фкые have [e;-, e*] - 0, for j = к this holds anyway. (iii) Suppose X = ekoX e (TMiP)k; then 0 = Liee.(o)(eb X) = [ey, X]-eko [es, X]. (iv) Suppose X e (Тм>р);, У e (TMiP)k, кф1ф j; then e,- о X = 0 and 0 = Liee/oX(o)(e,, Y) = e,- о Liex(o)(e,-, Y) = e,- о [X, et о У] - e,- о [X, e,-] о У - e,- о [X, У] о ef = -е,- о [X, Y]. 0 Claim (iv) shows that for any к the subbundle with germs of sections ©;Y*C7m,p) j is integrable. According to the Frobenius theorem (cf. for example [War, p. 41]) there is a (germ of a) submersion fk : (M, p) ->¦ (Cdim(r*M>', 0) such that the fibres are the integral manifolds of this subbundle. Then ф^ fk : (M, p) -*¦ (CdimM, 0) is an isomorphism. The submanifolds (Mk, p) := (@;#t /;)~40), p) yield the decomposition (M, p) = ni=i(^*- P) wi*
18 Definition and first properties of F-manifolds Claim: (v) If X, Y e prt-%4i/, then X о Y e pr^TMttP. (vi) If E is an Euler field then eko E € prk~xTMkP. Proof of the claim: (v) The product X о Y is contained in (9M>/) ¦ prk~ because this sheaf is multiplication invariant. Now X о Y e prk~lTMktP if and only if [Z, X о Y] e {TM,p)j for any j and any Z e {TMtP)y, but [Z, W]= Liez(X о Y) = UtejoZ{X о У) = ej о Liez(X о Y) e (TM,p)j. (vi) Analogously, for any Z € (TMiP)j -[Z, ekoE]= LieetoE(ej о Z) = Lieeto?(°)(e/, Z) + Lieeto?(e/) о Z + LieetO?-(Z) о е,- E{ej) oZ + UeetoE(Z) о е,- е (TM xTMk<p is true Claim (v) and (vi) show that the multiplication on (M, /?) and an Euler field E come from multiplication and vector fields on the submanifolds (M^, p) via the decomposition. These satisfy B.7) and B.8): this is just the restriction of B.7) and B.8) to pr^TMktP. ? Examples 2.12 (i) The manifold M = С with coordinate и and unit field e = ~- аи with multiplication e о e = e is an F-manifold. The field ? = u-e = M^-is an Euler field of weight 1. The space of allEuler fields of weight d is d ¦ E + С ¦ e. One has only to check B.7) and B.8) for X - Y = Z = W = e and compare B.10). Any 1-dimensional F-manifold is locally isomorphic to an open subset of this F-manifold (С, о, e). It will be called A\. (ii) From® and Proposition 2.10 one obtains the F-manifold A" — (С, о, e) with coordinates u\,...,un, idempotent vector fields e-, = j^, semisimple multiplication <?, о ej — 5,7e,-, unit field e = ?\ et and an Euler field E = J2t и; • et of weight 1. Because of Theorem 2.11, the space of Euler fields of weight d is d ¦ E + ?,. С • et. Also because of Theorem 2.11, any F-manifold M with semisimple multi- multiplication is locally isomorphic to an open subset of the F-manifold A". The induced local coordinates u\,..., un on M are unique up to renumbering and shift. They are called canonical coordinates, following Dubrovin. They are the eigenvalues of a locally defined Euler field of weight 1. 2.5 F-manifolds and potentiality 19 (iii) Any Frobenius manifold is an F-manifold [HM][Manl, I§5], see section 2.5. (iv) Especially, the complex orbit space of a finite Coxeter group carries the structure of a Frobenius manifold [Du2][Du3, Lecture 4]. The F-manifold structure will be discussed in section 5.3, the Frobenius manifold structure in section 5.4. Here we only give the multiplication and the Euler fields for the 2-dimensional F-manifolds him), m > 2, with /2B) = A2, /2C) = A2, /2D) = B2 = C2, /2E) =: Hi, /2F) = G2. The manifold is M — C2 with coordinates t\, t2; we denote 5,- := jf. Unit field e and multiplication о are given by e = Si and S2 о S2 = t2 ¦ S\. An Euler iield E of weight 1 is E = tiSi + ^t2S2. The space of global Euler fields of weight «5? is d ¦ E + С • e for m > 3. The caustic is 1С = {t e M \ t2 = 0} for m > 3 and 1С = 0 for m = 2. The multiplication is semisimple outside of 1С; the germ {M, t) is an irreducible germ of an F-manifold if and only if t € 1С. One can check all of this by hand. We will come back to it in Theorem 4.7, when more general results allow more insight. (v) Another 2-dimensional F-manifold is C2 with coordinates tt, t2, unit field e = 8] and multiplication о given by S2 о S2 = 0. Here all germs (M, t) are irreducible and isomorphic. E\ := fiSi is an Euler field of weight 1. Contrary to the above cases with generically semisimple multiplication, here the space of Euler fields of weight 0 is infinite dimensional, by B.8): {E | Lie?(o) = 0 • o} = [E | [S\, E] =0, o2 о \b2, t\ = 0} = {e,5, + s2(t2)S2 | e, б С, е2 б СЫС2), 5,(e2) = 0}. B.12) (vi) The base space of a semiuniversal unfolding of an isolated hypersurface singularity is (a germ of) an F-manifold. The multiplication was defined first by K. Saito [SK6, A.5)] [SK9, A.3)]. A good part of the geometry of F-manifolds that will be developed in the next sections is classical in the case of hypersurface singularities, from different points of view. We will discuss this in section 5.1. (vii) Also the base of a semiuniversal unfolding of a boundary singularity is (a germ of) an F-manifold, compare section 5.2. There are certainly more classes of semiuniversal unfoldings which carry the structure of F-manifolds. 2.5 F-manifolds and potentiality The integrability condition B.5) for the multiplication in F-manifolds and the potentiality condition in Frobenius manifolds are closely related. For semisimple multiplication this has been known previously (with Theorem 3.2
20 Definition and first properties of F-manifolds 2.5 F-manifolds and potentiality 21 (i) ¦<=>¦ (ii) and e.g. [Hi, Theorem 3.1]). Here we give a general version, requiring neither semisimple multiplication nor flatness of the metric. Remarks 2.13 (a) In this section we need some basic notions from differential geometry: connections, covariant derivative of vector fields, torsion freeness, metric, Levi-Civita connection. For the real C°°-case one finds these in any text- textbook on differential geometry. The translation to the complex and holomorphic case here is straightforward. (b) Let M be a manifold with a connection V. The covariant derivative VXT of a (k, /)-tensor with respect to a vector field is defined exactly as the Lie derivative Lie* Г in Remark 2.7, starting with the covariant derivatives of vector fields. The operator Vx is a derivation on the sheaf of {k, /)-tensors just as Lie*. But Vx is also OM-linear in X, opposite to Lie*. Therefore VT is a (k + 1,0-tensor. Theorem 2.14 Let (M, o, V) be a manifold M with a commutative and as- associative multiplication о on TM and with a torsion free connection V. By definition, V о (X, Y, Z) is symmetric in Y and Z. If the C, \)-tensor Vo is symmetric in all three arguments, then the multipli- multiplication satisfies for any local vector fields X and Y LieXoy(o) = X о Liey(o) + Y о Liex(o). B.13) Proof. The term V о (X, Y, Z) = VX(F о Z) - VX(Y) oZ -Yo VX(Z) is symmetric in Y and Z. The D, l)-tensor (X, Y, Z, W) h> V о (X, Y о Z, W) + W о V о (X, Y, Z) = VX(Y о Z о W) - VX(Y) oZoW -Yo VX(Z) oW -YoZo V* (IV) B.14) is symmetric in Y, Z, W. A simple calculation using the torsion freeness of V shows (LieXoK(o) - X о Liey(o) -Yo Liex(o))(Z, W) = Vo(XoY,Z,W)-XoVo(Y,Z,W)-YoVo(X,Z,W) -Vo(ZoW,X,Y) + ZoVo(W,X,Y)+WoVo(Z,X, Y). B.15) If V о is symmetric in all three arguments then V о (X о Y, Z, W) + Z о V о (W, X, Y) + W о V о (Z, X, Y) B.16) is symmetric in X, Y, Z, W because of the symmetry of the tensor in B.14). Then the right hand side of B.15) vanishes. ? Theorem 2.15 Let (M, o, e, g) be a manifold with a commutative and asso- associative multiplication о on TM, a unit field e, and a metric g (a symmetric nondegenerate bilinear form) on TM which is multiplication invariant, i.e. the Q,0)-tensor A, A(X, Y, Z) := g(X, Y о Z), B.17) is symmetric in all three arguments. V denotes the Levi-Civita connection of the metric. The coidentity e is the I—form which is defined by e(X) = g(X, e). The following conditions are equivalent: (i) The manifold with multiplication and unit (M, o, e) is an F-manifold and s is closed. (ii) The D, 0)-tensor VA is symmetric in all four arguments. (Hi) The C, \)-tensor Vo is symmetric in all three arguments. Proof. The Levi-Civita connection satisfies Vg = 0. Therefore VA(X, Y, Z, IV) = Xg(Y, Z о IV) - g(VxF, ZoW) -g(Y,WoVxZ)-g(Y,ZoVxW) = g(Y, VX(Z о IV) - W о VXZ - Z о VXW) = g(Y, V о (X, Z, IV)). B.18) The metric g is nondegenerate and VA(X, Y, Z, IV) is always symmetric in Y, Z, W. Equation B.18) shows (ii) <=» (iii). Because of Vg = 0 and the torsion freeness VXF — VyX = [X, Y], the 1-form s satisfies de(X, Y) = X(e(Y)) - Y(s(X)) - e([X, Y]) = g(Y, Vxe) - g(X, VYe) = -VA(X, Y, e, e) + VA(F, X, e, e). B.19) Hence (ii) =» de = 0; with Theorem 2.14 this gives (ii) =» (i). It remains to show (i) => (ii). The equations B.20) and B.21) follow from the definition of Vo and from B.18), V о (X, Y, e) = Y о V о (X, e, ё), VA(X, U, Y, e) = VA(X, U о Y, e, e). B.20) B.21)
22 Definition and first properties of F-manifolds One calculates with B.15) and B.18) g(e, (Uexovio) -Xo Liey(o) - У о Lie* (o))(Z, W)) = VA(X о Y, e, Z, W) - VA(Y, X, Z, W) - VA(X, Y, Z, W) - VA(Z о W, e, X, Y) + VA(W, Z, X, Y) + VA(Z, W, X, Y). B.22) If (i) holds then B.19), B.21), and B.22) imply VA(Y, X, Z, W) - VA(W, Z, X, Y) = -VA{X, Y, Z, W) + VA(Z, W, X, Y). B.23) The left hand side is symmetric in X and Z, the right hand side is skewsymmetric in X and Z, so both sides vanish. VA is symmetric in all four arguments. D Lemma 2.16 Let (M, g, V) be a manifold with metric g and Levi-Civita con- connection V. Then a vector field Z is flat, i.e. VZ = 0, if and only if Liez(g) = 0 and the 1-form ez := g(Z,.) is closed. Proof. The connection V is torsion free and satisfies Vg = 0. Therefore (cf. B.19)) dez(X, Y) = g{Y, VXZ) - g(X, Vy Z), , Y) = g(Y, VXZ) + g(X, VrZ). B.24) B.25) ? Remarks 2.17 (a) Let (M, o, e, g) satisfy the hypotheses of Theorem 2.15 and let g be flat. Then condition (ii) in Theorem 2.15 is equivalent to the existence of a local potential Ф 6 OM,P (for any p e M) with (ATZ)O = A(X, У, Z) for any flat local vector fields X, Y, Z. (b) In view of this the conditions (ii) and (iii) in Theorem 2.15 are called potentiality conditions. (c) The manifold (M, о, е, Е, g) is a Frobenius manifold if it satisfies the hypotheses and conditions in Theorem 2.15, if g is flat, if Liee(g) = 0 (respec- (respectively e is flat), and if ? is an Euler field (of weight 1, with respect to M as F-manifold), with UeE(g) = D ¦ g for some DeC (cf. Definition 9.1). Chapter 3 Massive F-manifolds and Lagrange maps In this section the relation between F-manifolds and symplectic geometry is discussed. The most crucial fact is shown in section 3.1: the analytic spectrum of a massive (i.e. with generically semisimple multiplication) F-manifold M is a Lagrange variety L С T*M; and a Lagrange variety L С Т*М in the cotangent bundle of a manifold M supplies the manifold M with the structure of an F-manifold if and only if the map a : Тм —*¦ л* Oi from C.1) is an isomorphism. The condition that this map a : TM -*¦ n*OL is an isomorphism is close to Givental's notion of a miniversal Lagrange map [Gi2, ch. 13]. In section 3.4 the correspondence between massive F-manifolds and Lagrange maps is rewritten using this notion. If E is an Euler field in a massive F-manifold M then the holomorphic func- function F := a~'(?) : L —>• С satisfies dF\Lreg = a\Lreg (here a is the canonical 1-form on T*M). But as L may have singularities, the global existence of E and of such a holomorphic function is not clear. This is discussed in section 3.2. Much weaker than the existence of ? is the existence of a continuous function F : L -> С which is holomorphic on Lreg withdF\Lreg = a\Lreg. This is called a generating function for the massive F-manifold. It gives rise to the three notions bifurcation diagram, Lyashko-Looijenga map, and discriminant. They turn out to be holomorphic even if F is only continuous along Lsmg. This is discussed in section 3.5 for F-manifolds and in section 3.3 more generally for Lagrange maps. A good reference for the basic notions from symplectic geometry which are used in this chapter (Lagrange variety, Lagrange fibration, Lagrange map, generating function) is [AGV1, ch. 18]. 3.1 Lagrange property of massive F-manifolds Consider an n-dimensional manifold (M, o, e) with commutative and asso- associative multiplication on the tangent bundle and with unit field e. Its analytic 23
24 Massive F-manifolds and Lagrange maps spectrum L :— Specan(T^) is a subvariety of the cotangent bundle T*M. The cotangent bundle carries a natural symplectic structure, given by the 2-form da. Here a is the canonical 1-form, which is written as a = ?(. y,df, in local coordinates t\ ,...,?„ for the base and dual coordinates y\,... ,yn for the fibres Um -*• (Лт-MhUT'M, щ i-> yi). The isomorphism a : TM -> я» OL from B.2) can be expressed with a by C.1) where X eTM and X is any lift of X to a neighbourhood of L in T*M. The values of the function a(X) on л~1(р) are the eigenvalues of Xo on T^Af; this follows from Theorem and Lemma 2.1. Definition 3.1. A manifold (M, o, e) with commutative and associative mul- multiplication on the tangent bundle and with unit field e is massive if the multipli- multiplication is generically semisimple. Then the set of points where the multiplication is not semisimple is empty or a hypersurface, which is the caustic 1С (Proposition 2.6). In the rest of the paper we will study the local structure of massive F-manifolds at points where the multiplication is not semisimple. It is known that the analytic spectrum of a massive Frobenius manifold is Lagrange (compare [Au] and the references cited there). Theorem 3.2 to- together with Theorem 2.15 make the relations between the different conditions transparent. Theorem 3.2 Let (M, o, e) be a massive n-dimensional manifold M. The analytic spectrum L = SpecanG]y) с Т*М is an everywhere reduced subvariety. The map л : L -*¦ M is finite and flat. It is a branched covering of degree n, branched above the caustic 1С. The following conditions are equiv- equivalent. (i) The manifold (M, o, e) is a massive F-manifold; (ii) At any semisimple point p e M — 1С, the idempotent vector fields e\,..., en б Тм,р commute. (Hi) The subvariety L С Т*М is a Lagrange variety, i.e. a\Lreg is closed. Proof. The variety L— ж~1AС) is smooth, and л : L — л~1AС) -*¦ M — /Cisa covering. The sheaf п„Оъ (= Тм) is a free 0M-module, so a Cohen-Macaulay Ом-module and a Cohen-Macaulay ring. Then L is Cohen-Macaulay and everywhere reduced, as it is reduced at generic points (cf. [Lo2, pp. 49-51] for the notion Cohen-Macaulay and details of these arguments). 3.1 Lagrange property of massive F-manifolds 25 It remains to show the equivalences (i) <==>¦ (ii) ¦?=> (iii). (i) => (ii) follows from Theorem 2.11 and has been discussed in Example 2.12 (ii). (ii) =>• (i) is clear because B.8) holds everywhere if it holds at generic points (in fact, one point suffices). (ii) => (iii) We fix canonical coordinates м,- with ^ — e,- on (M, p) for a point p 6 M — 1С and the dual coordinates x,- on the fibres of the cotangent bundle G^ _>. {пТ'м)*От'м, ej h> Xj). Then locally above (M, p) the analytic spec- spectrum L is in these coordinates L = {(xj, uj) | Xi. -I \-xn = l, П = {J{(Xj,Uj)\xJ=8ij}. C.2) 1=1 The 1-form a = J^Xjdu, is closed in L — л ]AС). This set is open and dense in Lreg, hence L is a Lagrange variety. (iii) => (ii) Above a small neighbourhood U of p e M - K, the analyticjpec- trum consists of и smooth components L*, к = 1,... ,n, with ж : L^ -—>¦ U. An idempotent vector field et can be lifted to vector fields e~i in neighbourhoods Uk of Lk in T*M such that they are tangent to all Lk. The commutator [el, e'j] is a lift of the commutator [et, ej] in these neighbourhoods Uk. el)) - de(ej, ej))\u , - e~\Lk(&ik) ~ = 0. But a : T M is an isomorphism, so [eit ej] =0 ? Theorem 3.3 (a) Let (M, o, e) be a massive F-manifold. A vector field E is an Euler field of weight с eC if and only if C.3) (b)Let(M, p)= ni=iWb p) be the decomposition ofthe germ of a massive F-manifold into irreducible germs of F-manifolds (Mk, o, ek). (i) The space of (germs of) Euler fields of weight Ofor (M, p) is the abelian Lie algebra ?;t=, С • ek.
26 Massive F-manifolds and Lagrange maps (ii) There is a unique continuous function F : (L, ж '(/>)) —*• (С, 0) on the multigerm (L, n~x(p)) which has value 0 on it~](p), is holomorphic on Lreg and satisfies (dF)\Lreg = a\Lreg. (Hi) An Eulerfield of weight с ф 0for(M, p) exists if'and onlyif"thisfunction F is holomorphic. In that case, с ¦ a~](F) is an Euler field of weight с and С • a~](F) + Yl'k=i"-" ' ek й the Lie algebra of all Euler fields on the germ (M, p). Proof. (a)Itis sufficientto prove this locallyforagerm(M, р)шй\ p e M—IC. This germ is isomorphic to A". A vector field E = YL"i=\ е<еь 8i 6 Ом,Р, is an Euler field of weight с if and only if ds,- = с ¦ d«,- (Theorem 2.11 and Example 2.12 (ii)). Going into the proof of 3.2 (ii) =>• (iii), one sees that this is equivalent to C.3). (b) The multigerm (L, тт~[(р)) has / components, and the space of locally constant functions on it has dimension I. The function (multigerm) F exists because a\Lreg is closed. This will be explained in the next section (Lemma 3.4). All statements follow now with (a). ? 3.2 Existence of Euler fields By Theorem 3.2, the analytic spectrum (L, A) of an irreducible germ (M, p) of a massive F-manifold is a germ of an (often singular) Lagrange variety, and (L, A) '-*¦ (T*M, X) —у (М, p) is a germ of a Lagrange map. The paper [Gi2] of Givental is devoted to such objects. It contains implicitly many results on massive F-manifolds. It will be extremely useful and often cited in the following. The question when does a germ of a massive F-manifold have an Euler field of weight 1 is reduced by Theorem 3.3 (b)(iii) to the question when is the function germ F holomorphic. Partial answers are given in Corollary 3.8 and Lemma 3.9. We start with a more general situation, as in [Gi2, chapter 1.1]. Let(L,0) С {CN ,G) be a reduced complex space germ. Statements on germs will often be formulated using representatives, but they are welldefined for the germs, e.g. '<*\Lng is closed' fora 6 QkCN 0. Lemma 3.4 Let a € ?2<J,,, 0 be closed on Lreg. Then there exists a unique function germ F : (L, 0) —> (C, 0) which is holomorphic on LKg, continuous on L and satisfies Proof: The germ (L, 0) is homeomorphic to a cone as it admits a Whitney stratification. One can integrate a along paths corresponding to such a cone 3.2 Existence of Euler fields 27 structure, starting from 0. One obtains a continuous function F on L, which is holomorphic on Lreg because of da\Lrcg = 0 and which satisfies dF\L =a\L . The unicity of F with value F@) = 0 is clear. ? Which germs (L, 0) have the property that all such function germs are holo- holomorphic on (L, 0)? This property has not been studied much. It can be seen to be in line with the normality and maximality of complex structures and is weaker than maximality. It can be rephrased as Hqjv((L, 0)) = 0. Here #glV((L, 0)) is the cohomology of the de Rham complex (?$„,„/{<» 6 nj.Wi0 | w\Lns = 0}, d), C.4) which is considered in [Gi2, chapter 1.1]. We state some known results on this cohomology. Theorem 3.5 (a) (Poincare-Lemma, [Gi2, chapter 1.1]) If (L, 0) is weighted homogeneous with positive weights then Hqjv((L, 0)) = 0. (b) ([Va5]) Suppose that (L, 0) is a germ of a hypersurface with an isolated singularity, (L, 0) = (/"'(О), 0) С (C"+1, 0) and f : (C"+1, 0) -> (C, 0) is a holomorphic function with an isolated singularity. Then C.5) -i- -dimOc,,,0/ f,^~ (с) (essentially Varchenko and Givental, [Gi2, chapter 1.2]) Let (L, 0) be as in (b) with ц - x # 0. The class [q] e H?iv((L, 0)) ofrj e ^?«+',0 '¦* not vanishing ifdrj is a volume form, i.e. dr) = hdxQ... dxn with h@) ф 0. Remarks 3.6 (i) The proofs of (b) and (c) use the GauB-Manin connection for isolated hypersurface singularities. (ii) Theorem 3.5 (c) was formulated in [Gi2, chapter 1.2] only for n = l.The missing piece of the proof for all n was the following fact, which at that time was only known for n = 1: The exponent of a form hdxo ... dxn is the minimal exponent if and only ifh(O) ф 0. This fact has been established by M. Saito [SM4, C.11)] for all n. (iii) By a result of K. Saito [SKI], an isolated hypersurface singularity (L, 0) = (/~'@), 0) С (C"+1, 0) is weighted homogeneous (with positive weights) if arid only if д — r = 0. (iv) For us only the case n = 1 in Theorem 3.5 (b) and (c) is relevant. Proposition 3.7, which is also due to Givental, implies the following:
28 Massive F-manifolds and Lagrange maps Of all isolated hypersurface singularities (Z,, 0) = (/"'@), 0) С (C+1, 0) only the curve singularities (n = 1) turn up as germs of Lagrange varieties. These are, of course, germs of Lagrange varieties with respect to any volume form on (C2, 0). (v) If (L, 0) С ((S, 0), со) is the germ of a Lagrange variety in a symplectic space S with symplectic form со, then the class [a] 6 H^iv((L, 0» of some a with da = со is independent of the choice of a. It is called the characteristic class of (L, 0) с ((S, 0), <и). (vi) Givental made the conjecture [Gi2, chapter 1.2]: Let (Z,, 0) be an n- dimensional Lagrange germ. IfH?iv((L, 0)) / 0 then H^iv{{L, 0)) # 0 and the characteristic class [a] e Hq!v{{L, 0)) is nonzero. It is true for n = 1 because of Theorem 3.5 and Remark 3.6 (iii). Givental sees the conjecture to be analogous to a conjecture of Arnold's which was proved by Gromov 1985 (cf. [Gi2, chapter 1.2]): any real closed Lagrange manifold L с T*W has a nonvanishing characteristic class [a] e Я'(?> R). Proposition 3.7 ([Gi2, chapter 1.1]) An n-dimensional germ (Z.,0) of a Lagrange variety with embedding dimension embdim (Z,, 0) = n + к is a product of a k-dimensional Lagrange germ (Z/, 0) with embdim (Z/, 0) = 2k and a smooth (n — k)-dimensional Lagrange germ (L", 0); here the decompo- decomposition of(L, 0) corresponds to a decomposition (E, 0), со) = (E', 0), со') x (E", 0), со") of the symplectic space germ (S, 0) Э (L, 0). C.6) Proof. If к < n then a holomorphic function f on S exists with smooth fibre /~'@) Э L. The Hamilton flow of this function / respects L and the fibres of /. The spaces of orbits in /~'@) and L give a symplectic space germ of dimension 2n —2 and in it a Lagrange germ (e.g. [AGV1, 18.2]). To obtain a decomposition as in C.6) one chooses a germ (E, 0) С E, 0) of а 2и — 1-dimensional submanifold E in S which is transversal to the Hamilton field H/ of /. There is a unique section v in (rS)^ with co(Hf, v) = 1 and аАТрЪ, v) = 0 for p e E. The shift If of v with the Hamilton flow of / forms together with H/ a 2-dimensional integrable distribution on S, because This distribution is everywhere complementary and orthogonal to the inte- integrable distribution whose integral manifolds are the intersections of the fibres of / with the shifts of E by the Hamilton flow of /. This yields a decomposition (S, 0) = (C2, 0) x (E П /~'@), 0). One can check that the symplectic form decomposes as required. If к < п — 1 one repeats this process. D 3.3 Lyashko-Looijenga maps and graphs of Lagrange maps 29 Corollary 3.8 (a) Let {L, 0) be an n-dimensional Lagrange germ isomorphic to (Z/, 0) x (C"~', 0) as complex space germ. Then (Z/, 0) is a plane cwve singularity. The characteristic class [a] 6 Hgiv((L,Q)) is vanishing if and only if{L', 0) is weighted homogeneous. (b) Let (L,X) С (Т*М, X) be the analytic spectrum of an irreducible germ (M, p) of a massive F-manifold. Suppose (L, X) = (Z/, 0) x (C", 0). Then there exists an Eulerfield of weight 1 on (M, p) if and only if(L', 0) is weighted homogeneous. Proof, (a) Proposition 3.7, Theorem 3.5, and Remark 3.6 (iii). (b) Part (a) and Theorem 3.3 (b)(iii). ? In Proposition 5.27 for any plane curve singularity (Z/, 0) irreducible germs of F-manifolds with analytic spectrum (Z,, X) = (Z/, 0) x (C"~l, 0) for some n will be constructed. So, often there exists no Euler field of weight 1 on a germ of a massive F-manifold. On the other hand, the Poincare-Lemma 3.5 (a) and Proposition 3.7 say that an Euler field of weight 1 exists on a germ of a massive F-manifold (M, p) if the multigerm (Z,, л ~' (/?)) of the analytic spectrum is at all points of n~l(p) a product of a smooth germ and a germ which is weighted homogeneous with positive weights. Also, we have the following. Lemma 3.9 Let M be a massive F-manifold and F : L -> С a continuous function with dF\Lreg = a\Lreg. Then a~l(F\(L - п~\1С))) is an Eulerfieldof weight 1 on M — 1С. It extends to an Euler field on M if(L, X) is at all points X € L outside of a subset of codimension > 2 a product of a smooth germ and a germ which is weighted homogeneous with positive weights. Proof. Suppose К С L is a subset of codimension > 2 with this property. Then F is holomorphic in L - К because of the Poincare-Lemma 3.5 (a) and Proposition 3.7. The Euler field extends to M - я(К). But jt(K) also has codimension > 2. So the Euler field extends to M (and F is holomorphic on L). О 3.3 Lyashko-Looijenga maps and graphs of Lagrange maps In this section classical facts on Lagrange maps are presented, close to [Gi2, chapter 1.3], but slightly more general. They will be used in sections 3.4-4.1. LetZ, с Т*М be a Lagrange variety (not necessarily smooth) in the cotangent bundle of an /n-dimensional connected manifold M. We assume:
30 Massive F-manifolds and Lagrange maps (a) The projection ж : L —>• M is abranched covering of degree n, that is, there exists a subvariety tfcAf suchthatrc : L-n~\K) -> М- К is a covering of degree n (ж : L —> M is not necessarily flat). (b) There exists a. generating function F : L —*¦ C, that is, a continuous func- function which is holomorphic on Lng with &F\L = a\L (locally such a function exists by Lemma 3.4). Such a function F will be fixed. It can be considered as a multivalued function on M - A"; the 1-graph of this multivalued function is L - n~[(K). The Lyashko-Looijenga map A = (Л|,..., Л„) : M -»¦ С" of L С Т*М and F is defined as follows: for q e M - K, the roots of the unitary polynomial z" + 5Z"=i л;(?)г"~' are the values of F on n~\q). It extends to a holomorphic map on M because F is holomorphic on Lreg and continuous on L. The reduced Lyashko-Looijenga map A(red) = (A(*d) Л^'°) : M -»• С of L С Г*М and F is defined as follows: for q e M -V, the roots {fed) =2 A{fed){q)zn~' are the values of F on ,(я) = ± Z!Xejr-i(?) F(A.). It also extends of the unitary polynomial z" + n~\q), shifted by their centre -± ?) to a holomorphic map on M. Its significance will be discussed after Remarks 3.11. The front <!>?. of L С T*M and F is the image Im(F, pr) с С х М of (F, pr) : L -> С x M. It is the zero set of the polynomial z" + YTi=\ л< ' z"~'. So, it is an analytic hypersurface even if F is not holomorphic on all ofL. Following Teissier ([Te2, 2.4, 5.5], [Lo2, 4.C]), the development ф2 с РГ*(С х М) of this hypersurface Ot in С х М is defined as the closure in PT*(C x M) of the set of tangent hyperplanes at the smooth points of Ф/.. It is an analytic subvariety and a Legendre variety with respect to the canonical contact structure on РГ*(С х М). The map С x T*M -» P7"(C x M), (c,A.)i-> ((dz-A.)-'@),(c, p)) C.7) (A. € г;м and dz - A. € 7-{*>p)(C x M) S ГС*С х г;М) identifies СхГМ with the open subset in РГ*(С х М) of hyperplanes in the tangent spaces which do not contain С ¦ ^. The induced contact structure on С х T*M is given by the 1-form dz — a. The following fact is well known. It is one way in which the relation between Lagrange and Legendre maps can be made explicit (e.g. [AGV1, ch. 18-20]). To check it, one has to consider F as a multivalued function on M - К and Ф/, as its graph. 3.3 Lyashko-Looijenga maps and graphs of Lagrange maps 31 Proposition 3.10 The embedding С х Т*М ^- РГ*(С х М) identifies the graphlm(F, id) С С x T*M of F : L ->• С with the development Ф^ of the front Ф?. Remarks 3.11 It has some nontrivial consequences. (i) The development Ф^ is contained in the open subset of P7*(C x M) of hyperplanes in the tangent spaces which do not contain ^. Therefore the vector field -fz is everywhere transversal to the front Ф^.. (ii) The polynomial z"+5Z"=i л< z"~'has no multiple factors and the branched covering Фх. -»¦ M has degree n: over any point p € M — К the varieties L and Фь have n points and the points of Ф^ have n tangent planes; so, also Ф?, has n points over a generic point p e M — K. (iii) The graph Im(F, id) = Ф?. is an analytic variety even if F is not holo- holomorphic on all of L. (iv) The composition of maps Ф^ —-*¦ Im(F, id) —> L is a bijective mor- phism. It is an isomorphism if and only if F is holomorphic on L. Also, the continuous map L —> Ф^ is a morphism if and only if F is holomorphic on L. (v) The Lagrange variety L <ZT*M together with the values of F at one point of each connectivity component of L and any of the following data determine each otheruniquely: the front Ф?., the development Ф?,, the Lyashko-Looijenga map Л, the generating function F as a multivalued function on the base M. To motivate the reduced Lyashko-Looijenga map, we have to talk about Lagrange maps and their isomorphisms ([AGV1, ch. 18], [Gi2, 3.1]). A Lagrange map is a diagram L '->¦ (S, со) —*¦ M where L is a Lagrange variety in a symplectic manifold E, со) and S -> M is a Lagrange fibration. An isomorphism between two Lagrange maps is given by an isomorphism of the Lagrange fibrations which maps one Lagrange variety to the other. An automorphism of T*M -> M as a Lagrange fibration which fixes the base is given by a shift in the fibres, T*M^T*M, Xh+X + dS, where S : M -> С is holomorphic ([AGV1, 18.5]). So, regarding T*M -> M as a Lagrange fibration means to forget the 0-section and the 1-form a, but to keep the Lagrange fibration and the class a + {dS | S : M -> С holomorphic} of 1-forms. Corollary 3.12 Let L c-> T*M —>¦ M be as above (satisfying the assumptions (a) and (b)) with I connectivity components and points X\, ... ,Xi, one in each connectivity component. The data in (i)-(iii) are equivalent.
32 Massive F-manifolds and Lagrange maps 3.4 Miniversal Lagrange maps and F-manifolds 33 (i) The diagram Lc-^-T*M^-Masa Lagrange map and the differences F(ki) - F(kj) e С of values of F, (ii) the generating function F modulo addition of a function on the base, (Hi) the reduced Lyashko-Looijenga map A(reJ) : M -*¦ С". Proof. (i)=> (ii): Integrating the l-formsina+{d5 | S : M -*¦ С holomorphic} gives (ii). (ii) =4- (iii) Definition of A(red). (iii) => (i): The map @, Л(гп/)) = @, A2red\ ..., A<rerf>) : M -»• C" is the Lyashko-Looijenga map of a Lagrange variety in T*M which differs from the original Lagrange variety only by the shift of d(^-Ai) in the fibres. The map @, A(rerf)) determines this Lagrange variety and a generating function for it because of Remark 3.11 (v). Then (i) is obtained. ? 3.4 Miniversal Lagrange maps and F-manifolds The notion of a miniversal germ of a Lagrange map is central in Givental's paper [Gi2]. We need a slight generalization to multigerms, taking a semilocal viewpoint. Let LcTMbeaLagrange variety with finite projection ж : L -*¦ M. The germ at the base point p e M of L =->• T*M -*¦ M is the diagram (L, jt-'Cp)) «-»¦ (T*M, T*pM) -> (M, p). Here (L, л-'(р)) is a multigerm. The map of germs (T*M, T*M) -»• (M, p) is the cotangent bundle of the germ (M, p); it is a germ in the base, but not in the fibre. For this diagram the morphisms O M,p and С 0 (c, X) i- -> 3t.(Ol)p. (c. X) н> (с + e(X))|L „М -> x*(OL)p/mp ¦ ir.(OL)p, (c +a(X))|7r*(OL)p/m;, • jr,@L)p C.8) C.9) are welldefined. Here X is in both cases a lift of X to 7 *M. These morphisms are not invariants of the diagram as a germ at the base point p e M of a Lagrange map because the identification of the Lagrange fibration with the cotangent bundle of (M, p) is unique only up to shifts in the fibres and only the class of 1-forms a + [dS \ S : M -*¦ С holomorphic} is uniquely determined (cf. section 3.3). But being an isomorphism or epimorphism in C.8) and C.9) is clearly a property of the germ at p of the Lagrange map. Definition 3.13 The germ atpeMofi^ T*M ->• M as a Lagrange map is called miniversal (versal) if the morphism in C.9) is an isomorphism (epimorphism) (cf. [Gi2, chapter 1.3]). We are only interested in the case of a flat projection ж : L -*¦ M.Wellknown criteria of flatness for finite maps (cf. [Fi, 3.13]) give the next lemma. Lemma 3.14 The following conditions are equivalent. (i) Thegermat p e MofL =->• T*M -*¦ M as a Lagrange map is miniversal with flat projection ж : (L, n~x(p)) -*¦ (M, p), (ii) it is miniversal with deg ж = 1 + dim M, (iii) the morphism in C.8) is an isomorphism, (iv) the Lagrange map is miniversal at all points in a neighbourhood of p&M. Example 3.15 A miniversal germ at a base point of a Lagrange map with a projection ж : L -»• M which is not flat is given by the germ at 0 e C2 of the Lagrange fibration C4 ->• С2, (уг.уз.Ь^з) •->¦ (t2, b) with со = dy2dt2 + dy^db and by the Lagrange variety L which is the union of two appropriate planes and which is defined by the ideal (Л. Уз) n(y2-t2,y3- h) = (y2, уъ) ¦ (Уг -h,y3- h). C.10) Now let M be a massive n-dimensional F-manifold with analytic spectrum L С T*M. Then (L, n~\p)) ^ (T*M, T*M) -+ (M, p) is for any p e M a versal, but not a miniversal germ at the base point peMofa Lagrange map. But there is a miniversal one. The germ of the fibration at p whose fibres are the orbits of the unit field e is denoted by pre : (M, p) -*¦ (M(r), pw). The fibrewise linear function on T*M which corresponds to e is called y\. Its Hamilton field ? := Hyi is a lift of e to T*M. It leaves the hypersurface ^,"'A) С Т*М and the Lagrange variety L С yf 'A) invariant. The orbits of ~e in yj"'(l) form a germ of a 2n — 2-dimensional symplectic manifold with a Lagrange fibration, which can be identified with the cotangent bundle
34 Massive F-manifolds and Lagrange maps 3.5 Lyashko-Looijenga map of an F-manifold 35 But this identification is only unique up to shifts in the fibres. The orbits of 7 in L form a Lagrange variety L(r) С T*Mir). The germ at p(r) e Mir) of the diagram L(r) <^-> T*M^r) -> M(r) is unique up to isomorphism of germs in the base of Lagrange maps. It will be called the restricted Lagrange map of the germ (M, p) of the F-manifold M. An explicit description will be given in the proof of the next result. Theorem 3.16 (a) The restricted Lagrange map of the germ(M, p) of a massive F-manifold is miniversal with flat projection n^ : L(r) -> Л/(г). (b) It determines the germ (M, p) of the F-manifold uniquely. (c) Any miniversal germ at a base point of a Lagrange map L' =-> T*M' -> M' with flat projection L' —*¦ M' is the restricted Lagrange map of a germ of a massive F-manifold. Proof: (a) In order to be as explicit as possible we choose coordinates t — (t\,t') = (t\,...,tn):(M,p)^ (C", 0) with e(t\) = 1. The dual coordinates on T*M are (yi,..., у„) = (уь у') = у. The multiplication is given by ^ о ;?- = J^k afj(t')~ and the analytic spectrum L is I ? \(y,t) e C" x (C,0) | y, = 1, yiyj = ?4(r')yJ. C.11) The restricted Lagrange map is represented by the Lagrange fibration С x (C-'.O)-* (C"-',0), (y',t')h+ t' C.12) with canonical 1-form a' := ?/>2 У'1^' ап<^ by the Lagrange variety {(/, t') 6 1,0) | yiyj j = 4(г') + ^<а')Л for i, j > 2J = L(r). <fc>2 J C.13) The equations for the Lagrange variety in C.13) show that the morphism in C.8) for this Lagrange map with fixed canonical 1-form a' is an isomorphism. This implies that the restricted Lagrange map for (Af, p) is miniversal with flat projection. (b) and (c) Any miniversal germ at a base point p' e M' of a Lagrange map L' °-> T*M' —*¦ M' with flat projection л' : L' —*¦ M' can be represented by a Lagrange fibration as in C.12) and a Lagrange variety as in C.13). Defining L by C.11) and M := С x (M', p') and e := ^-, one obtains an F-manifold with unit field e and analytic spectrum L. It remains to show that this does not depend on the way in which the Lagrange fibration is identified with the cotangent bundle of (C"""!, 0) in C.12). But one sees easily that a shift in the fibres of C.12) of the type y,- i-> y,- + ff for some holomorphic function 5 : (C"',0) -> С on the base corresponds only to a change of the coordinate fields ^-,..., ^- and the coordinate t\ in M and thus to a shift of the section {0} x (M\ p') in M -> (M1, p'). It does not affect L and the multiplication on (M, p). D Let (M, p) be a germ of a massive F-manifold. The germ дрю := [X 6 TM,p I [e, X] = 0} C.14) is a free 0WM]/7ir)-module of rank n. It is an Ом"\р^-algebra because of Liee(o) = 0 • o. The functions a(X) for X e Qpm are invariant with respect to ? and induce holomorphic functions on L(r). One obtains a map Lemma 3.17 The map a(r) is an isomorphism of ОMi,i ^-algebras. C.15) Proof. The isomorphism a: Тм,р -*¦ {ix^Odp maps the e-invariant vector fields in (M, p) to the (T-invariant functions in GГ*0/.)р. ? This isomorphism a(r) is closely related to C.8) for the restricted Lagrange map of (M, p): An isomorphism as in C.8) requires the choice of a 1-form for its Lagrange fibration. The choice of a function t\ : (M, p) -» (C, 0) with e(t\) = 1 yields such a 1-form: the 1-form which is induced by a — dt\ (a — dt\ on T*M is «^-invariant and vanishes on "e and induces a 1-form on the space of «Г-orbits of The choice of t\ also yields an isomorphism ОМ(-),ри Ф TM(o,pM -> CV'>,,<" • e Ф {X e ?PM I X(ti) = 0} = дрю. C.16) One sees with the proof of Theorem 3.16 (a) that the composition of this iso- isomorphism with a(r) gives the isomorphism in C.8) for the restricted Lagrange map of (M, p) (the germ {M, p) in C.8) in this case is (M(r), p(r))). 3.5 Lyashko-Looijenga map of an F-manifold Definition 3.18. Let (M, o, e) be a massive n-dimensional F-manifold with analytic spectrum L С Т*М.
36 Massive F-manifolds and Lagrange maps (a) A generating function F for (M, o, e) is a generating function for L, that is, a continuous function F : L —> С which is holomorphic on Lreg with dF|tl4 =«!*„. (b) Let F be a generating function for (M, o, e). (i) The bifurcation diagram В с M of(M,o,e, F)isthe set of points p e M such that F has less than л different values on я (p). (ii) TheLyashko-Looijengamap A = (Ль ..., Л„): M -*¦ С oi(M, о, е, F) is the Lyashko-Looijenga map of F as the generating function for L С T*M (cf. section 3.3). (iii) The discriminant V С M of (А/, о, г, F) is V := Л~'@). The discriminant will be discussed in section 4.1. A generating function for an F-manifold exists locally (Lemma 3.4), but not necessarily globally. A holomorphic generating function F corresponds to an Euler field E := a~l(F) of weight 1 (Theorem 3.3); then the values of F on л~1(р), р е М, are the eigenvalues of Eo : TpM —*¦ TpM. The objects B, A, V of Definition 3.18 (b) are welldefined for (M, o, e, E) if E is such an Euler field. The restriction of л : L —>¦ M to the complement of the caustic /C is a covering 7Г : n~[(M — /C) -*¦ M -K. of degree n, and ж'1 (М -К,) is smooth. Hence a generating function F is holomorphic on ж~1(М — /C) and corresponds to an Euler field E on M — /C. Results and examples about the extendability of E to M are given in Lemma 3.9 and Theorem 5.30. The bifurcation diagram В of (M, o, e, F) contains the caustic /C. The caustic is a hypersurface or empty (Proposition 2.6) and invariant with respect to the unit field e (Remark 2.9 (vii)). The bifurcation diagram has the same properties: the restriction of F to an open set U С M — К with canonical coordinates (mj ,..., «„) corresponds to an Euler field E = ?(«, + ri)ei f°r some r,- 6 C, and the bifurcation diagram is the hypersurface В П U = U П [и | и,- + r,- = My + ry for some i # /}. It is invariant with respect to e because of е(и,- — м;) = 0. The Lyashko-Looijenga map for the F-manifold A" = (С", о, e) (Example 2.12 (ii)) with Euler field E = Y, «;«•; and Euler field-function F := a(E) is л(«): С" -> С", « h> ((-1L-(")); C.17) here o\ (и), ...,ап{и) are the symmetric polynomials. The group of automor- automorphisms of the F-manifold A" which respect the Euler field E is the sym- symmetric group Sn which permutes the coordinates u\,... ,и„. The map Л(п) 3.5 Lyashko-Looijenga map of an F-manifold 37 is the quotient map for this group. It is branched along the bifurcation diagram В = [и | и; = uj for some i # j). The image Л(л)(#) is the hypersurface Vм := la e C" | z" + J] a;z"~' has multiple roots} С С". C.18) The restriction Л(п) : С" — В ->¦ С - Vм induces an F-manifold structure on ?n _ ?>«, with unit field e(n) ._ dA<«)(e) _ _„ and Euler field K-17- C19) oat C.20) This F-manifold (C - ?>(n), o, e(n)) will be denoted by A\/Sn. Theorem 3.19 Let (M, o,e)bea massive F-manifold with generating function F : L —*¦ С and Lyashko-Looijenga map A : M —>¦ C. 77ien Л-'ф(п)) = В anddA(e) = e(n). Г/ie restriction A : M - В ->¦ С - P(n) и аи immersion and locally an isomorphism of F-manifolds. It maps the Euler field a1 (F\M-b) on M - В to the Euler field Ew. Proof In M — /C the multiplication is semisimple and locally the values of the generating function are canonical coordinates. The map Л factors on M — В locally into an isomorphism to A" and into the map Л(п). ? The most important part of Theorem 3.19 is that A : M - В-*¦ С - V(n) is locally biholomorphic. The following statements for germs will also be useful. Lemma 3.20 Let (M, p) = I~L=i(^b P) be я germ of a massive F-manifold with analytic spectrum Land with decomposition into irreducible germs (M^, p) of dimension n^, Ylnk — n- (a) There exists precisely one generating function on the multigerm (L, 7t~l(p))forany choice of its values on 7r~'(p) = {k\,..., Л;}. (b) Choose a function t\ : (M, p) ->¦ С with e(t\) = 1. The values of a generating function for the points in L above an orbit of e are of the form t\ +a constant. The entry Aj of a Lyashko-Looijenga map A : (M, p) —*¦ С is a polynomial of degree i in t\ with coefficients in {g e Ом p \ e(g) = 0} andleading coefficient (-1УС).
38 Massive F-manifolds and Lagrange maps (c) Choose representatives Mi for the germs (Mk, p)and Lyashko—Looijenga maps A[k] : Mk -»¦ C". Then the function A = (A,,..., Л„): П* Mk -> С which is defined by 1=1 ,?-' = f] [z" C.21) is a Lyashko-Looijenga map for the representative fj Mk of the germ (M, p). Any Lyashko—Looijenga map for (M, p) is of this type. Proof, (a) Lemma 3.4. (b) It suffices to prove the first part for an orbit of e in M — JC. There the generating function comes from an Euler field. The formulas B.9) and B.10) imply Liee(?o) = id. The values of F are the eigenvalues of Eo. (c) The map A[k] corresponds to an Euler field ?№1 (at least) on Mk - BMt. The sum J2k ?W is an Euler field on TlWk - BMk) by Proposition 2.10. The corresponding generating function extends to f] Mk and has the given Л as its Lyashko-Looijenga map. The last statement follows with (a). D Consider the projection pre : (M, p) —*¦ (M(r\ p^) whose fibres are the orbits of e (section 3.4). The e-invariant hypersurfaces В and /C project to hypersurfaces in M^r\ which are called the restricted bifurcation diagram 23(r) and the restricted caustic /C(r). In section 3.3 the restricted Lagrange map was defined as the germ at p(r> e M(r) of a Lagrange map L(r) ^ T*M(r) -»¦ M(r\ Because of Corollary 3.12 the notion of a reduced Lyashko-Looijenga map is welldefined for the restricted Lagrange map (independently of the identification of the Lagrange fibration with the cotangent bundle of M(r)). The space of orbits of the field eM (formula C.19)) in C" can be identified with (a e C" | fli = 0) = {0} x С = С" and is equipped with the co- coordinate system (fl2,..., an) = a'. The projection to this orbit space is denoted by pr(n) : С -»¦ С1, and the image of ?>(n) is _,):= l чл-l i=2 has multiple roots | C.22) (it is isomorphic to the discriminant of the singularity or F-manifold An-\, cf. section 5.1). Corollary 3.21 Let (M, p)be the germ of a massive F-manifold, F a generating function, A (ASndr>) the (reduced) Lyashko-Looijenga map of(M, o, e, F). The map A^^ : (M,p) -*¦ C is constant along the orbits of e. The induced map A(rerfXl") : (M(r\ p(r)) -*¦ C" is a reduced Lyashko-Looijenga 3.5 Lyashko-Looijenga map of an F-manifold 39 ap for the restricted Lagrange map. The following diagram commutes, the 'diagonal morphism is Л"*" = pr<"> о Л = Л"*"™ о pre, т> (М,р) С" (M{r\ Pir)) > С' 1И-1 C.23) The restriction is locally biholomorphic. - X>(An_i) C.24) Proof. The map A(red) is constant along the orbits of e because of Lemma 3.20 (b). The formulas C.11), C.12), C.13) show that A<re</)(r) is a reduced Lyashko-Looijenga map for the restricted Lagrange map. The rest follows from Theorem 3.19. П
Chapter 4 Discriminants and modality of F-manifolds 4.1 Discriminant of an F-manifold 41 Discriminants play a central role in singularity theory. Usually they have a rich geometry and say a lot about the mappings or other objects from which they are derived. The discriminant V of a massive F-manifold M with a generating function (cf. Definition 3.18) is an excellent model case of such discriminants, having many typical properties. Together with the unit field it determines the whole F-manifold in a nice geometric way. This is discussed in section 4.1 (cf. Corollary 4.6). In section 4.3 results from singularity theory are adapted to show that the discriminant and also the bifurcation diagram are free divisors under certain hypotheses. The classification of germs of 2-dimensional massive F-manifolds is nice and is carried out in section 4.2. Already for 3-dimensional germs it is vast (cf. section 5.5). In section 4.4 the Lyashko-Looijenga map is used to prove that the automorphism group of a germ of a massive F-manifold is finite. There also the notions modality and /^-constant stratum from singularity theory are adapted to F-manifolds. In section 4.5 the relation between analytic spectrum and multiplication is generalized. This allows F-manifolds to be found in natural geometric situations (e.g. hypersurface and boundary singularities) and to be written down in an economic way (e.g. in 5.22, 5.27, 5.30, 5.32). 4.1 Discriminant of an F-manifold Let (M, o, e, F) be a massive n-dimensional F-manifold with a generating function F : L -> С and Lyashko-Looijenga map Л = (ЛЬ...,Л„):М-» С"; the discriminant of (M, o, e, F) is the hypersurface V = Л~'@) с М (Definition 3.18). If F is holomorphic and E = a~'(F) is its Euler field then Л„ = (-1)" • det(F,o), and the discriminant is the set of points where the multiplication with E is not invertible. By the definition of Ли, the discriminant is V = n(F ' @))/Theorem 4.1 will give an isomorphism between F @) and the development V С VT*M of V. We need an identification of subsets of T*M and FT*M. The fibrewise linear function on T*M which corresponds to e is called y\. The canonical map 3>j~'(l)—у ?Т*М D.1) identifies yf'A) С Т*М with the open subset in VT*M of hyperplanes in the tangent spaces of M which do not contain С • e. The restriction to yf'O) of the canonical 1-form a on T*M gives the contact structure on з>Г'A) which is induced by the canonical contact structure on Theorem 4.1 Let(M, o, e, F)andF~\G) С L С y^d) С Т* М be as above. The canonical map y^1 A) <-» PT*M identifies F~l@) with the development V С РТ*М of the discriminant T>. Proof. We want to make use of the discussion of fronts and graphs (section 3.3) and of the restricted Lagrange map (section 3.4). It is sufficient to consider the germ (M, p) for some p 6 V. We choose coordinates (t\,..., tn) and (yi, ¦ • •. Уп) as in the proof of Theorem 3.16. The generating function F for I <^. j*M -> M takes the form F(y', t) = h + F(r)(/, t'), ' D.2) where Fw is a generating function of L(r) with respect to C.13), C.12), and a' = Хл>2 yi&i- The isomorphism with a sign (-«i. pre) :(«,?)->(Cx Mw, 0 x p(r>) D.3) maps the discriminant to the front Im(FCr), я(г)) of Lw and F(r). The development of this front is identified with the graph Im(F(r), id) С С x T*Mir) of F(r> : L(r) -> C, by Proposition 3.10 and the embedding (cf. formula C.7)) С x T*Mir) <-> РГ*(С x M(r)), (-fi, X) ^ ((df, + X)-'(O), ( - tu ttw(X))). D.4) But D.1), D.3), and D.4) together also yield an isomorphism yj~'A) -> С х Г*М(Г>, which maps F~' @) to this graph. D Remarks 4.2 (i) The most important consequence is that V С FT*M does not contain a hyperplane which contains С • e (cf. Remark 3.11 (i)). Therefore the unit field e is everywhere transversal to the discriminant V. 40
42 Discriminants and modality of F-manifolds (ii) In the proof only the choice of t\ is essential. It is equivalent to various other choices: the choice of a section of the projection (M, p) —*¦ (MM, p^r)), the choice of a 1-form a' for the Lagrange fibration in the restricted Lagrange map. (iii) Fixing such a choice of t\, one obtains together with a' and F(r) a Lyashko-Looijenga map A(r) : (M(r\ p{r)) -» С" for F{r). Then D.3) iden- identifies the entry Л„ of the Lyashko-Looijenga map Л : (M, p) -*¦ C" with the polynomial (-f,)" + Y!U A,(r)(-f,)"-'. (iv) The set F~l@) С L is not an analytic hypersurface of L at points of LSing where F is not holomorphic. But Theorem 4.1 shows that it is everywhere a subvariety of L of pure codimension 1. Examples where F is not holomorphic will be given in section 5.5. (v)Let(M, p) = ni=i(^b p)bethedecompositionofagerm(M, p)intoir- p)intoirreducible germs of F-manifolds. A Lyashko-Looijenga map Л : (M, p) -*¦ С corresponds to Lyashko-Looijenga maps A'*' : {Mk, p) —*¦ C* for the irre- irreducible germs, in a way which was described in Lemma 3.20 (c). One obtains especially An = ]~It Aj*], and the germ (V, p) of the discriminant is . P) = \J <k UMJ-p D.5) the union of products of smooth germs with the discriminants for the irreducible germs (of course, it is possible that ((Aj*])~'(O), p) = 0 for some or all k). (vi) The development V of the discriminant V gives the tangent hyperplanes to V. Let (M, p) = Y\'k^i(Mk, p) be as in (iv), and л~\р) = {Xu ..., A,}. Theorem4.1 says that the tangenthyperplanes to (?>, /?)arethosehyperplanes A^'@) С ТрМ for which F(A*) = 0. Especially, if/ = 1 and F(A.,) = 0, then Aj"'(O) С TPM is the nilpotent subalgebra of ГРМ and the unique tangent hyperplane to (V, p). The general case fits with Lemma 2.1 (iv) and D.5). (vii)The equality a\Lirg = dF|Lres shows immediately that F~'@) С yf 'A) is a Legendre subvariety. (viii) If Л € F~'@) П T*M, there is a canonical projection from the Leg- Legendre germ (JF~' @), X) to one component of the Lagrange multigerm (L(r\ n(r) (pM)). it is a bijective morphism. It is an isomorphism if and only if F is holomorphic at (L, A.) (see Remark 3.11 (iv)). One can recover the multiplication on a massive n-dimensional F-manifold M from the unit field e and a discriminant V if the orbits of e are sufficiently large. To make this precise, we introduce the following notion. 4.1 Discriminant of an F-manifold 43 Definition 4.3 A massive F-manifold (M, o, e, F) with generating function F is in standard form if there exists globally a projection pre : M —>• M(r) to a manifold M(" such that (a) the fibres are the orbits of e (and thus connected), ф) they are with their affine linear structure isomorphic to an open (connected) subset of C, (y) the projection pre : T> —> M^ is a branched covering of degree n. Remarks 4.4 (i) If (M, o, e, F) is a massive F-manifold with generating func- function F and properties (a) and (/3) then 1 ( A,,prJ V n ) M С x M(r) D.6) is an embedding because of e{—\i^\) = 1 (Lemma 3.20 (b)). The F-manifold M can be extended uniquely to an F-manifold isomorphic to С x M(r). Also the generating function F can be extended. The discriminant of this extended F-manifold satisfies (y) because of V = A~'@) and Lemma 3.20 (b). (ii) For (M, o, e, F) as in (i) the coordinate t\ :— — ^A\ is distinguished and, up to the addition of a constant, even independent of the choice of F. Nevertheless it does not seem to have good properties: In the case of the simplest 3-dimensional irreducible germs of F-manifolds, A3, B3, Щ, it is not part of the coordinate system of a nice normal form (section 5.3). Using the data in [Du2] one can also check that — ^Ai is not a fiat coordinate of the Frobenius manifolds A3, B3, Я3. Corollary 4.5 Let (M, o, e, F) be a massive F-manifold with generating func- function F and in standard form. (a) The branch locus of the branched covering pre : V -*¦ M(r' is Vsing, the set pre(DSins) of critical values is the restricted bifurcation diagram B^ = pre(B). (b) The union of the shifts of F~\0) with the Hamilton field 7 = Hyi is the analytic spectrum L С Т*М. (c) The data (M, o, e, F) and {M, e, V) are equivalent. Proof, (a) Theorem 4.1 implies that all tangent hyperplanes to T> are transversal to the unit field. Therefore the branch locus is only T>smg. (b)and(c)ThediscriminantI?andtheunitfieldedetermineF~l(O) С yf'(l) because of Theorem 4.1. The uni on of shifts of F ~' @) with the Hamilton field "e = Hyi is finite of degree n over M because of (y) and it is contained in L,
44 Discriminants and modality of F-manifolds so it is L. The function F : L -*¦ С is determined by F~'@) С L and by the linearity of F along the orbits of e". ? Theorem 4.1 and Corollary 4.5 (b) give one way to recover the multiplication of a massive F-manifold (A/, o, e, F) in standard form from the discriminant T> and the unit field e. The following is a more elementary way. Corollary 4.6 Let (M, o, e, F) be a massive n-dimensional F-manifold with generating function F and in standard form. The multiplication can be recover- recovered from the discriminant T> and the unit field e in the following way. The multipli- multiplication is semisimple outside of the bifurcation diagram В = pr~l(pre(Vsing)). For a point p € M — B, the idempotent vectors е,{р) е TpM with е,{р) о ejip) — Sjjeiip) are uniquely determined by (i) and (ii): (i) the unit vector is e{p) = 2^"=1 ei(p), (ii) the multigerm (P, T> П pr~\pre(p))) has exactly n tangent hyperplanes; their shifts to TpM with e are the hyperplanes ф,^ С • e,- С ТрМ, к = 1.....П. Proof. Remark 4.2 (vi). D 4.2 2-dimensional F-manifolds The only 1-dimensional germ of an F-manifold is A\ (Example 2.12 (i)). The class of 3-dimensional germs of massive F-manifolds is already vast. Examples and a partial classification will be given in section 5.5. But the classification of 2-dimensional germs of F-manifolds is nice. Theorem 4.7 (a) The only germs of 2-dimensional massive F-manifolds are, up to isomorphism, the germs him), m e N>2, with I2B) = A2, I2C) = A2, /2D) = Въ /2E) =: H2, /2F) = G2,from Example 2.12 (iv): The multiplication on (M, p) = (C2, 0) with coordinates t\, t2 and Si := ^ is given by e := <5| and S2 о S2 = t™~2 ¦ S\. An Euler field of weight 1 is E = /,Si+ ^t2&2. Its discriminant is V = {t\t\-^t% — 0}. The germ I2(m) of an F-manifold is irreducible for m > 3 with caustic and bifurcation diagram /С = В = {t 112 = 0}. The space of Euler fields of weight d is d ¦ E + С • e for m>3. (b) The only germ of a 2-dimensional not massive F-manifold is the germ (C2, 0) from Example 2.12 (v) with multiplication given by e := S\ andS2o82 = 0. The caustic is empty. An Euler field of weight 1 is E = t[&\. The space of all Euler fields of weight 0 is {e,5| + e2(t2)B2 \ e{ e C, e2(t2) e C{t2}}. 4.2 2-dimensional F-manifolds 45 Proof (a) Givental [Gi2, 1.3, p. 3253] classified the 1-dimensional miniversal germs of Lagrange maps with flat projection. Together with Theorem 3.16 this yields implicitly the classification of the 2-dimensional irreducible germs of massive F-manifolds. But we can recover this in a simple way and we need to be more explicit. Let (M, p) be a 2-dimensional germ of a massive F-manifold with projection pre : (M, p)->- (Af(r), p(r))to the space of orbits of e. There isaunique genera- generating function F : (L, л~1(р)) -*¦ (С, 0). Its bifurcation diagram В с М and restricted bifurcation diagram B(r) С A/(r)arethehypersurfacesS = pr~^(p<r^) By Corollary 3.21, the reduced Lyashko-Looijenga map Л*""*^: (M(r\ p(r)) —*¦ (C, 0) of the restricted Lagrange map is a cyclic branched cov- covering of some order in, and the Lyashko-Looijenga map Л : M —*¦ C2 is also a cyclic branched covering of order in, branched along B. Because of Corollary 3.12 and Theorem 3.16, this branching order in of д(г"/)(г) determines the germ {M, p) of the F-manifold up to isomorphism. It remains to determine the allowed in and explicit formulas for the F-manifolds. Now consider the manifolds C2 with multiplication on ГС2 given by e = Si and S2oS2 = t™~2 ¦ h\. The analytic spectrum L С Т*С2 is L - {(yi, y2, D-7) It is an exercise to see that L is a Lagrange variety with generating function F = t\ + ^y2h with respect to a — y\dti+y2dt2, i.e. onehas(a— dF)\Lnsg = 0. Then this gives an F-manifold and E = a~' (F) — t\S\ + ^t2b2 is an Euler field of weight 1. The Lyashko-Looijenga map Л is Л : С2 -+ С2, (г,, t2) t+(-2tut2- ^ (- D.8) It is branched along В = {t \ t2 = 0} of degree m. So I2(m) is the desired F-manifold for any branching order m = m > 2. The same calculation yields for m = m = 1 the F-manifold A1/S2 on C2 — {t 112 = 0} (section 3.5) with meromorphic multiplication along {t \t2 = 0}. (b) Let (M, p) = (C2, 0) be the germ of a 2-dimensional not massive F- manifoldwithe — <5i.ThenLiee(o) = 0and[e, S2] = OimplyLiee(<52 0<52) — 0. Hence <52 о <$2 = p{t2)8i + y(h)&2 for some 0(f2), yUi) e C{t2}. The field S2 := S2 - {y(t2)Si satisfies S2 о S2 = (fi(t2) + \y(t2f)8{ and [<$,, S2] = 0. Changing coordinates we may suppose S2 = S2, y(t2) = 0. The analytic spectrum is L = {{y\, y2% t\, t2) | y\ = 1, y2 • y2 = fi{t2)}. The F-manifold is not massive, hence fi(t2) = 0. One checks with B.7) and B.8)
46 Discriminants and modality of F-manifolds easily that this multiplication gives an F-manifold and that the space of Euler fields is as claimed. ? Let us discuss the role which 2-dimensional germs of F-manifolds can play for higher dimensional massive F-manifolds. The set a € С | z" + ]Г) ац"~1 has a root of multiplicity > 3 J С Vм с С" D.9) is an algebraic subvariety of C" of codimension 2 (see the proof of Proposition 2.5). Given a massive F-manifold M, the space /CC) := {p € M | P(TpM) >-C,l l)}ciCcM D.10) of points p such that (M, p) does not decompose into 1- and 2-dimensional germs of F-manifolds is empty or an analytic subvariety (Propostion 2.5). Theorem 4.8 Let (M, o, e) be a massive F-manifold with generating func- function F. (a) The function F is holomorphic on л~х{М — /C<3)) and gives rise to an Euler field of weight 1 on M — KP\ (b) //codim/CC) > 2 then F is holomorphic on L and E = a~\F) is an Euler field of weight 1 on M. (c) One has /CC) С A^"'3'), and A-'CD("'3)) -/C<3> is analytic of pure codimension 2. Thus codim/CC) > 2 «=>¦ codimA^"'3') > 2. (d) The restriction of the Lyashko—Looijenga map A : M - A-'(D("-3)) -» C" -D("'3) is locally a branched covering, branched along В — A (X>("'3)). IfpeB- A-'fD'"'3') and A(p) e Х>(л) - D("-3) ore smootfi pomta o/fAe hypersurfaces В andV^ and if there the branching order is m, then (M, p) is the germ of an F-manifold of type h(m) x A"~2. Proof: (a) and (b) Each germ (L, Xk) of the analytic spectrum (L, n~](p)) = (L, {Ль ..., Л/}) of a reducible germ (M, p) = Ylk(Mk, P) is the product of a smooth germ with the analytic spectrum of (M^, p). The analytic spectrum of him) (m > 2) is isomorphic to (C, 0) x ({y2, t2) | y\ = t™^2}, 0). One applies Lemma 3.9. (c) A Lyashko-Looijenga map of h (m) is a cyclic branched covering of order m, branched along the bifurcation diagram. This together with Lemma 3.20 (c) 4.3 Logarithmic vector fields 47 implies that locally around a point p € M — /CC) the fibres of the Lyashko- Looijenga map A : M —*¦ C" are finite. Therefore codim/v/(A~'(I>('l-3)), p) = codime(D(;l-3))=2. (d) The map Л determines the multiplication of the F-manifold M (Theorem 3.19). One uses this, Lemma 3.20 (c) and properties of h(m). Q Many interesting F-manifolds, e.g. those forhypersurface singularities, boun- boundary singularities, finite Coxeter groups (sections 5.1, 5.2, 5.3), satisfy the property codim/CC) > 2 and have an Euler field of weight 1. 4.3 Logarithmic vector fields K. Saito [SK4] introduced the notions of logarithmic vector fields and free divisors. Let H С M be a reduced hypersurface in an я-dimensional manifold M. The sheaf Der^ (log H) С Тм oflogarithmic vector fields consists of those holomorphic vector fields which are tangent to Hreg. This sheaf is discussed in detail in section 8.1. There it is shown that it is a coherent and reflexive Ом-niodule. The hypersurface H is a free divisor if Der^(log H) is a free См-module of rank n. The results in this section are not really new. They had been established in various generality by Bruce [Bru], Givental [Gi2, chapter 1.4]), Lyashko [Lyl][Ly3], K. Saito [SK6][SK9], Terao [Тег], and Zakalyukin [Za] as results for hypersurface singularities, boundary singularities or miniversal Lagrange maps. But the formulation using the multiplication of F-manifolds is especially nice. Theorem 4.9 Let(M, o, e) be a massive F-manifold with Euler field E of weight 1, generating function F — a(?) and discriminant V = (det(?o))~'@) = (a) The discriminant is a free divisor with Der^(log V) = E о Тм- (b) The kernel of the map 0) D-11) is ker nv = EoTM= DerM(log V). D.12) Proof, (a) The sheaf E о Тм is a free 0M-module of rank n. Therefore (a) follows from D.12).
48 Discriminants and modality of F-manifolds (b) The Од*-module я,С/г-1(о) has support V. Equation D.12) holds in M - 23. The set V П В = Vsing (cf. Corollary 4.5 (a)) has codimension 2 in M. The following shows that it is sufficient to prove D.12) in V - Vsing. Let M — Vsing <-+ M be the inclusion. The Riemann extension theorem says OM — i*(OM-vlins) (cf. for example [Fi, 2.23]). Now E oTM satisfies EoTM = /»(? о TM\M^t>siJ D.13) because it is a free CM-module. The sheaves kerap and DerA/(log2?) satisfy the analogous equations because of their definition. Hence D.12) holds in M if it holds in V - Vsing. Let p 6 V—Vsing. We choose a small neighbourhood U of p with canonical coordinates u\,...,un centred at p, with РП U = {и | u{ =0} and with Euler field E = и\в\ + ?,>2(«; + л>; for some г,- е С — {0}. With the notation of the proof of Theorem 3.2 (ii) =Ф- (iii) we have a = J2 *,d«,-, F~\0) П n~4U) = {(x, «) | xj = Su, щ = 0}, and for any vector field X = X! ?/*i € T^(U) Therefore = f,@, иг,..., «„)¦ i=2 = ? о TW)/, = DerWiP(log23). D.14) П Remark 4.10 One can see Theorem 4.9 (a) in a different way: there is a criterion of K. Saito [SK4, Lemma A.9)]. To apply it, one has to show [E°TM,EoTM]cEoTM. D.15) With B.5) and B.6) one calculates for any two (local) vector fields X, Y [EoX,EoY] = Eo ([X, EoY]-[Y,E6X]-Eo[X, Y]). D.16) In the rest of this section (M, о, е, Е) will be a massive F-manifold which is equipped with an Euler field E of weight 1 and which is in standard form (Definition 4.3). The map pre : M -*¦ M(r) is the projection to the space of orbits of e. The sheaf of e-invariant vector fields Q := {X € (preXTM | [e, X] = 0} D.17) 49 4.3 Logarithmic vector fields is a free O^o-module of rank n. Because of Liee(o) = 0 it is also an CVo- algebra. Theorem 4.11 Let (M, o, e, E) be a massive F-manifold with Euler field E of weight 1 and in standard form, (a) (рге)*Тм = G® (pre)*{E о Тм). (b) The kernel of the map (pre)*av : (pre\TM -»¦ (pre о ^),OF-i@), X н is (pre)*{E о Тм). The restriction (pre)**v ¦ Q -> (pre ° Jr).CV-'@) is an isomorphism of О м^-algebras. D.18) D.19) D.20) Proof, (a) It follows from (b). (b) The kernel of (ргг)»а© is (pre)t(E о Тм) because of Theorem 4.9 (b). The F-manifold in standard form has a global restricted Lagrange map L^r) c-* T*M(r) -»¦ M(r) (the identification of its Lagrange fibration with T*Mir) -»• M(r) is unique only up to shifts in the fibres). The canonical projection F~l@) -*¦ L(r) is bijective (Corollary 4.5 (b)), and then an isomorphism because F is holomorphic. It induces an isomor- isomorphism (pre о ir),Of-i(o) = (тг<г))*О^(г). The composition with D.20) is the isomorphism a<r> : Q from Lemma 3.17. D.21) D Theorem 4.9 and Theorem 4.11 are translations to F-manifolds of state- statements in [SK6, A.6)] [SK9, A.7)] for hypersurface singularities. In fact, K. Saito essentially used D.20) to define the multiplication on Q for hypersurface singularities. The arguments in Lemma 4.12 and Theorem 4.13 are due to Lyashko [Lyl] [Ly3] and Terao [Ter], see also Bruce [Bru]. Again(M, o, e, ?')isamassive F-manifold with Euler field of weight 1 and in standard form. We choose a function t\:M~>C with e(t\) = 1 (e.g. t\ = — ? Л i, cf. Lemma 3.20 (b)). This choice simplifies the formulation of the results in Lemma 4.12. The vector fields in TMn will be identified with their (unique)
50 Discriminants and modality of F-manifolds lifts in {X e Q | Х{Ц) = 0} CG С (рге)„Тм. The projection to TMcr, of all possible lifts to M of vector fields in Mu' is D.22) Lemma 4.12 Let (M, о, е, Е, t\) be as above. (a) n-l 0 -t\-e D.23) (bj ?acft vector field in M(r) wft/cft /г/to to a vector field in (pre)*DerM (log V) lifts to a unique vector field in D.23). (c) The vector fields in M(r) which lift to vector fields in (pre)*DerM(logX>) are tangent to the restricted bifurcation diagram B(r) = pre(B) С M(r) and form the free 0Mm-module of rank n - 1 n-l d(pr,)((r,e - ?H*) с Z)erMW(logB(r)). D.24) k> 1 : (a) One sees inductively by multiplication with txe - E that for any t\e -{he- E)°k 6 (prt),(E о Тм) D.25) holds. The inclusion he — E e ? and Liee(o) = 0 imply (t\e - ?)°* 6 therefore л-l *=o tf e = D.26) Now the decomposition D.18) yields D.23). (b) The map pre : V -+ M(r) is a branched covering of degree n (Definition 4.3), so D.27) *=o is an isomorphism. Therefore any lift h ¦ e + X, h e (pre), OM, of X б Тмм can be replaced by a unique lift h ¦ e + X with h 6 0^Zq Omw ff and (/г - А)Ь = 0. If /г • e + X is tangent to D, then (ft - A)b = 0 is necessary and sufficient for ft • e + X to be tangent to P. Logarithmic vector fields 51 (с) A generic point p(r) 6 (B(r))reg has a preimage p e {Vsing)reg such that the projection of germs pre : (Vsing, p) -* (B('\ p(r)) is an isomorphism (Corollary 4.5 (a)). A vector field ft ¦ e + X, X e ТМю, which is tangent to Vreg is also tangent to (T>Smg)reg- Then X is tangent to (B{r))reg. One obtains the generators in D.24) by projection to TMm of the generators in D.23). ? The set /CC) С /С С M is the set of points p 6 M such that (M, p) does not decompose into 1- and 2-dimensional germs of F-manifolds (section 4.2). Theorem 4.13 Let (M, o, e, E) be a massive F-manifold with Eulerfield E of weight 1 and in standard form. Suppose that codim/CC) > 2. Tlien the restricted bifurcation diagram B(r) is a free divisor and D.24) is an equality. Proof. In view of Lemma 4.12 (c) it is sufficient to show that any vector field tangent to B(r) lifts to a vector field tangent to V. The projection pre :V—B^>- M(r) —B(r) is a covering of degree л. For any vector field X 6 TMw there exists a unique function hx 6 (рге)*От>-в such that A • e + X is tangent to V - В if and only if ft Ь-в = hx. One has to show that hx extends to a function in {pre\Ov if Xs DerMM(logB(r)). Then the unique lift ft • e + X with h б ф^ OMv)t\ and h\v = hx is tangent to V. Let p be a point in the set {p 6 B?^)^ | p(r) 6 (B(r))reg, pre : CD, p) -». (M<r\ p^) has degree 2}. D.28) Then the germ (D, p) is the product of (C"~2, 0) and the discriminant of the germ of an F-manifold of type him) (m > 2) (Remark 4.2 (v)). One can find coordinates (t\ ,t') = (tu..., tn) around p e M such that (D, p) С (М, p) ->• (M(r), p(r)) corresponds to ({(fi, О I rf - f2m = 0}, 0) С (С", 0) -> (С1, 0), г н> г'. D.29) Then (B(r), p{r)) = ({f' | Г2 = 0), 0). Obviously the vector fields tangent to (B(r), p(r)) locally have lifts to vector fields tangent to (D, p). The function ft* of a field X 6 DerMM(logB(r)) extends holomorphically to the set in D.28). The complement in V oiVreg —V — B and of the set in D.28) has codimension > 2 because of codim/CC) > 2. Therefore hx 6 {pre\Ov. D
52 Discriminants and modality ofF-manifolds 4.4 Isomorphisms and modality of germs of F-manifoIds The following three results are applications of Theorem 3.19 for the Lyashko- Looijengamap.Theywill be proved together. The tupIe((M, p), o, e, A)denotes the germ of an F-manifold with the function germ Л : (M, p) ->¦ C" as addi- additional structure. A map germ cp : (M, p) -> (M, p) respects Л if Л о ср = A. Theorem 4.14 The automorphism group of a germ (M, p) of a massive F- manifold is finite. Theorem 4.15 Let(M,o,e, F)bea massive F-manifold with generating func- function F and Lyashko-Looijenga map A : M ->• C". For any p\ e M the set (qeM\ ((M, pi), o, e, A) S ((M, q), o, e. A)} is discrete and closed in M. Corollary 4.16 Let (M,o,e,E)bea massive F-manifold with Eulerfield E of weight 1. For any p\ e M the set {qeM\ ((M, p,), o, e, E) S ((M, q), о, е, Е)} is discrete and closed in M. Proof. Corollary 4.16 follows from Theorem 4.15. For Theorem 4.14, it suffices to regard an irreducible germ of a massive F-manifold. The automorphisms of an irreducible germ respect a given Lyashko-Looijenga map Л because of Lemma 3.20 (a). So we may fix for Theorem 4.14 and Theorem 4.15 a massive F-manifold (M, o, e) and a Lyashko-Looijenga map Л : M —> C. The set Д := {(P. p')eMxM\ A(p) = Л(р')} has a reduced complex structure. It is a subset of(M-B)x(M-B)UBxS and the intersection А П (M - B) x (M - B) is smooth of dimension n. This follows from Theorem 3.19. ч Now consider an isomorphism <p : ((M, p), o, e, A) -» ((M, p'), o, e, A). The graph germ (G(<p), (p, p')) := ({(?, cp(q)) e M x M | q near p}, (p, p')) is a smooth analytic germ of dimension n and is contained in the germ (Д, (p, p')). It meets А П (M - B) x (M - B). Because of the purity of the dimension of an irreducible analytic germ, it is an irreducible component of 4.4 Isomorphisms and modality of germs of F-manifolds 53 the analytic germ (A, (p, p')). One can recover the map germ <p from the graph gSrm(G(<p), (p, p'))- The germ (A, (p, p')) consists offinitely many irreducible components. The case p = p' together with the remarks at the beginning of the proof give Theorem 4.14. For Theorem 4.15, we assume that there is an infinite sequence (p,-, ?>,-),¦ ещ of different points p-, e M and map germs <Pi : ((M, Pl), о, е, А) Д ((M, pi), o, e. A) and one accumulation point рж e M. The set Д - В х В is analytic of pure dimension n. It contains the germs (G(<Pi), (p\, Pi)) and the point (pi, p^). We can choose a suitable open neighbourhood U of (p\, poo) in M x M and a stratification (J Sa = U П A-BxB oft/ПА — В х В which consists of finitely many disjoint smooth connected constructible sets Sa and satisfies the boundary condition: The boundary Sa - Sa of a stratum Sa is a union of other strata. The germ (G(<Pi), (p\, pi)) is an n-dimensional irreducible component of the n-dimensional germ (A — В х В, (pi, pi)). There is a unique n-dimensional stratum whose closure contains (G(<Pi), (pi, Pi))- If (Pi.Pi) € Sa then this together with the boundary condition implies (Sa, (pi, pi)) С (<?fe)> (pi, Pi))- The germ (G(<Pi), (pi, Pi)) is the graph of the isomorphism щ. Therefore it intersects the germ ({p,} x M, (pb pi)) only in (pb pi); the same holds for Now there exists at least one stratum Sao which contains infinitely many of the points (pi, pi). The intersection of the analytic sets Sao and U П ({pi} x M) contains these points as isolated points. This is impossible. The above assumption was wrong. ? In singularity theory there are the notions of/x-constant stratum and (proper) modality of an isolated hypersurface singularity. One can define versions of them for the germ (M, p) of an F-manifold (M, o, e) (massive or not massive): The ix-constant stratum EД, p) is the analytic germ of points q e M such that the eigenspace decompositions of TqM and TpM have the same partition (cf. Proposition 2.5). Theidempotent fields e\,..., et of the decomposition (M, p) = \\k=\(Mk, p) into irreducible germs of F-manifolds commute and satisfy Lie,., (o) = 0 • o. So the germs (M, q) of points q in one integral manifold of e\,..., ei are isomorphic. This motivates the definition of the modality: modM(M, p) := dimEM, p) -1. D.30)
54 Discriminants and modality of F-manifolds Let (Sjf1, p) denote the /x-constant stratum of (Mk, p)\ Then Theorem 2.11 implies mod (M, p) — j) and modM(Mb p). D.31) D.32) For massive F-manifolds, Theorem 4.15 and Lemma 3.20 give more informa- information: Corollary 4.17 Let (M,p)= Y[[=i(Mk, P) be the germ of a massive F- manifold and A : (M, p) —*¦ С a Lyashko-Looijenga map. (a) There exist a representative 5M of the fi-constant stratum EД, р), а neighbourhood U С С о/О and an isomorphism U D.33) such that ifr~l([q] x U) is the integral manifold of eu ..., e/ which contains q. Any subset of points in 5Д П Л (A(p)) with isomorphic germs of F-manifolds is discrete and closed, (b) modM(M, p) = dimEM П A~\A(p)), p), D.34) sup(modM(M, q) \ q near p) = dim(A~x(A(p)), p). D.35) Proo/i (a) For / = 1, the existence of ф follows from the e-invariance of 5M and from e{-\hx) = 1 (Lemma 3.20 (b)). For arbitrary I, one uses D.31) and Lemma 3.20 (c): the maps A and (A[1],..., Ли) have the same germs of fibres, especially 5Д П A-\A(p)) = Y\ S™ П AW"'(Aw(p)). D.36) A germ (M, q) has only a finite number of Lyashko-Looijenga maps with fixed value at q (Lemma 3.20 (a)). The finiteness statement in Corollary 4.17 (a) follows from this and Theorem 4.15. (b) Equation D.34) follows from (a). A representative of the germ (A~l(A(p)), p) is stratified into constructible subsets which consist of the points q with the same partition for the eigenspace decomposition of TqM (Proposition 2.5). A point q e A~'(A(p)) in an open stratum with maximal dimension satisfies modM(M, q) = A~\A(p)), q) = &m(A-\A{p)), q) = dim(A-1(A(p)), p). D.37) 4.4 Isomorphisms and modality of germs of F-manifolds 55 This shows sup(modM(M, q)\qe A~\A(p)) near p) = dim(A-'(A(p)), p). D.38) The upper semicontinuity of the fibre dimension of A gives D.35). ? Remark 4.18 Gabrielov [Ga] proved in the case of isolated hypersurface sin- singularities the upper semicontinuity of the modality, modM(M, q) < modM(M, p) for q near p D.39) (and the equality with another version of modality which was defined by Arnold). He used D.34), D.35), and a result of himself, Lazzeri, and Le, which, translated to the F-manifold of a singularity (section 5.1), says: EД П A-'(A(p)), p) = (Л-ЧЛ(р)), р). D.40) The inequality D.39) is an immediate consequence of D.34), D.35) and D.40). But for other F-manifolds D.40) and D.39) are not clear. In the case of the simple hypersurface singularities, the base of the semiuni- versal unfolding is an F-manifold M = C" and the map A : M-B ->¦ <Cn-Vw is a finite covering. Therefore the complement M — В is а К (л, 1) space and the fundamental group is a subgroup of finite index of the braid group Br(n). This is the application of Looijenga [Lol] and Lyashko [Arl] of the map A, which led to the name Lyashko-Looijenga map. It can be generalized to F-manifolds. We call a massive F-manifold M simple if modM(M, p) = 0 for all p e M. This fits with the notions of simple hyper- hypersurface singularities, simple boundary singularities, and simple Lagrange maps ([Gi2, 1.3, p. 3251]). A distinguished class of simple F-manifolds are the F-manifolds of the finite Coxeter groups (section 5.3 and [Lol][Arl][Lyl][Ly3][Gi2]). There are other examples (Proposition 5.32 and Remark 5.33). A Lyashko-Looijenga map of a massive F-manifold is locally a branched covering if and only if M is simple (D.35) and Theorem 3.19). A detailed proof of the following result had been given by Looijenga [Lol, Theorem 2.1] (cf. also [Gi2, 1.4, Theorem 5]). Theorem 4.19 Let (M, p) — (C, p) be the germ of a simple F-manifold with fixed coordinates. Then, ife < sq for some sq, the space {z € C" | \z\ < e) — В is а К (л, 1) space. Its fundamental group is a subgroup of finite index of the braid group Br(n).
56 Discriminants and modality of F-manifolds 4.5 Analytic spectrum embedded differently The analytic spectrum L С T*M of an F-manifold determines the multiplica- multiplication on TM via the isomorphism (C.1) and B.2)) a : T M X н> a(x)\L. D.41) One can generalize this and replace L, T*M, and a by other spaces and other 1-forms. This allows F-manifolds to be found in natural geometric situations and to be encoded in an economic way. Corollary 4.21 and Definition 4.23 are the two most interesting special cases of Theorem 4.20. Theorem 4.20 Let the following data be given: manifolds Z and M, where M is connected and n-dimensional; a surjective map nz '¦ Z -> M which is everywhere a submersion; an everywhere n-dimensional reduced subvariety С С Z such that the restric- restriction jtc '¦ С -> M is finite; a I—form az on Z with the property: any local lift X e Tz of the zero vector field 0 6 TM satisfies az(X)\c = 0. D.42) Then (a) The map &c : TM X м- az(X)\c D.43) is welldefined; here X 6 Tz is any lift of X to a neighbourhood of С in Z. (b) The image L С T*Mofthe map q : С -> Т*М, г н* q(z) = (Хн ac(X)(z)) 6 T*ciz)M D.44) is a (reduced) variety. The map q : С -> L is a finite map, the projections ж : L —> M and tvq — ж о q are branched coverings. The composition of the maps q : n*OL -> {пс\Ос and a : TM X н>- a\X)\L D.45) is Ac = q о a. All three are OM-module homomorphisms. (c) The 1-forms a and az satisfy (q*a)|Cres = «zlc^- Therefore L is a Lagrange variety if and only ifaz\cns is exact. (d) The map a : 7д/ -> ntOi is an isomorphism if and only if (i) the map ac is injective, 4.5 Analytic spectrum embedded differently 57 С (jtc)*Oc is multiplication invariant, contains the unit \q 6 (лс)*Сс- (ii) its image Лс (Hi) the image Я In this case Ac '¦ 1м -> (псХ^с induces a (commutative and associative and) generically semisimple multiplication on Тм with global unit field and with analytic spectrum L. (e)Themap&c '¦ TM -> (лс)*0с provides Mwith the structure oj'a massive F-manifold if and only ifotzlc^ " exact and the conditions (i)-(iii) in (d) are satisfied. Proof, (a) This follows from D.42). (b) The equality dim С — n — dim M and nc finite imply that nc is open. M is connected, thus же is a branched covering. Using local coordinates for M and T*M one sees that q : С -> Т*М is an analytic map. The equality jtc = я о q is clear and shows that q is finite. Then L = q(C) is a variety and it is a branched covering. The equality ac = q о a follows from the definition of q. (c) There is an open subset M@) С M with analytic complement M - M@) such that л-cW') С С and n~\Mi0)) С L are smooth, nc : п^ Af@)and;r : л~\Мт) -> Af@) are coverings and q : Tr^W') -> is a covering on each component of л~1(М^). Now &c = q о a implies (d) The map q : я*О^ -> {жс)*Ос is an injective homomorphism of OM- algebras. If a : TM -> ж%Оь is an isomorphism then (i)-(iii) are obviously satisfied. Suppose that (i)-(iii) are satisfied. Then a : Тм -> ж*О^ is injective with multiplication invariant image a(TM) С ж+О^ and with \L 6 &(TM). The maps a and ac induce the same (commutative and associative) multiplication with global unit field on TM. We have to show that this multiplication is generically semisimple with an- analytic spectrum L. Then a : TM -> ж^Оь is an isomorphism and the proof of (d) is complete. If for each/? 6 M the linear forms in ж~1(р) С Т*М would generate a sub- space of T*M of dimension < n then a would not be injective. So, for a generic point p 6 M there exist n elements in ж~\р) С Т*М which form a basis of T*M. We claim that я~'(р) contains no elements other than these: ж~1(р) does not contain 0 6 T*M because of Ц 6 a(T^). From the multiplication invariance of аGд/) one derives easily that ж~х(р) does not contain any further elements. This extends to a small neighbourhood U of the generic point p 6 M: ж~1A]) consists of n sheets which form a basis of sections of T*M\ the map
58 Discriminants and modality of F-manifolds л\ц :Tu -*¦ л*(п \U)) is an isomorphism and induces a semisimple multi- multiplication on TM with analytic spectrum n~\U). Then L is the analytic spectrum of the multiplication on TM because M is connected. (e) By (c) and (d) and Theorem 3.2. D In Theorem 4.20 the map я : L -> M has degree n, but nc : С ->¦ A/ can have degree > n; and even if пс : С ->• A/ has degree л the map q : С -*¦ L does not need to be an isomorphism. Examples will be discussed below (Examples 4.24, Lemma 5.17). But the most important special case is the following. Corollary 4.21 LetZ, M, nz, С С Z, az, ac, L, andqbeasin Theorem4.20. Suppose that az\Creg is exact and &c '¦ TM -> {itc\Oc is an isomorphism. Then q : С -*¦ L is an isomorphism and ac=qoa provides M with the structure of a massive F-manifold with analytic spectrum L. Proof. Theorem 4.20 (e) gives all of the corollary except for the isomorphism q : С —*¦ L. This follows from the isomorphism q : n^Oi —у (ттс)*Ос and a universal property of the analytic spectrum. ? One can encode an irreducible germ of a massive F-manifold with data as in Corollary 4.21 such that the dimension of Z is minimal. Lemma 4.22 Let (A/, p) be an irreducible germ of a massive n-dimensional F-manifold. LetmC TpM denote the maximal ideal in TpM. (a) Then dim Z >n + dim m/m2 for any data as in Corollary 4.21. (b) There exist data as in Corollary 4.21 for (A/, p) with dimZ = n + dim m/m2 (the construction will be given in the proof). Proof, (a) Ttcl(p) — 7tzl(p) П С consists of one fat point with structure ring TPM. Its embedding dimension dim m/m2 is bounded by the dimension dim 7rJ' (p) = dim Z — n of the smooth fibre п^\(р). (b) One can choose coordinates (fb ..., tn) = (h, t') = t for (A/, p) with e = ^ as usual and with &-Ш= m с TPM and — | = т' с ТрМ D.46) D.47) 4.5 Analytic spectrum embedded differently 59 for m = 1+ dim m/m2. The dual coordinates on (T*M, T*M) are y{,..., yn, the analytic spectrum is (cf. B.1)) L = [(y, t)\yx= 1, yiyj = Y^afjb'bk}- D.48) Because of D.47) there exist functions bt e C{t'}[y2,..., ym] with yi\L=bi(y2,...,ym,t')\L fori=m + l, ...,л. D.49) We identify (A/, p) and (C\ 0) using (f,,..., *„) and define (Z, 0) = (С'" х C", 0). The embedding i: (Z, 0) = (Cm~l x C\ 0) ^ T*M, D.50) (Л,..., *т_ь О н»- ()>, 0 = A, xi,..., xm-Ubm+l(x, t'),.... bn(x, t'), t) provides canonical choices for the other data, 7TZ : (Z, 0) -»• (A/, p), (x, t) м- t, D.51) С =1-41-), D.52) 1=2 /=m+l The conditions in Corollary 4.21 are obviously satisfied. D.53) ? The notion of a generating family for a Lagrange map ([AGV1, ch. 19], [Gi2, 1.4]) motivates us to single out another special case of Theorem 4.20. Definition 4.23 Let Z,M,nz,C,az, and ac be as in Theorem 4.20 with ссг\с„д exact and ac : TM -*¦ (яс)*Ос injective with multiplication invari- invariant image ас(Тм) Э {lcb These data yield a massive F-manifold (A/, o, e). A function F : Z —*¦ С is a generating family for this F-manifold if az =dF and if С is the critical set of the map (F, nz): Z -*¦ С х A/. There are two reasons for the name generating family: A) The function F is considered as a family of functions on the fibres n^ip), p e M. B) The restriction of F to С is the lift of a generating function F : L -*¦ C, i.e. F = F о q; so the 1-graph of F as a multivalued function on M is L. In the case of a generating family the conditions D.42) and az exact are obvious. The most difficult condition is the multiplication invariance of It is not clear whether for any massive F-manifold M data (Z, nz, F) as in Definition 4.23 exist. But even many nonisomorphic data often exist. We illustrate this for the 2-dimensional germs him) of F-manifolds (section 4.1).
60 Discriminants and modality of F-manifolds Examples 4.24 Always (Z, 0) = (C x C2, 0) and (M, p) = (C2, 0) with pro- projection nz : (Z, 0) ->¦ (M, 0), (x, и, t2) н> (r,, t2) and e := S, := f, S2 := 3 (a) С = {(x, t) | xm-2 - t\ = 0}, az = drx + xd?2. These are data as in Corollary 4.21 for I2{m). (b) Generating family F =t{+ f*(t2 - u2)kdu (k > 1), С = {(*, r) 112 - x2 = 0},az|c=dF|c=(dri +c-x2*-1dr2)lcforsomec 6 С - {0}, ac(S2) ¦ These are data as in Definition 4.23 for I2Bk + 1), the map жс : С ->¦ Af has degree 2, the map q : С ->¦ L is the normalization and the maximalization of L (cf. [Fi, 2.26 and 2.29] for these notions). (c) Generating family F = tx +xk+]t2 - Щх*+2 (к > 1), С = {(x, t) \ (t2 - lc. These are data as in Definition 4.23 for I2Bk + 4), the map жс : С -+ M has degree 2, the map q : С -+ L is the maximalization of L (for the missing case /2D) compare Lemma 5.17). (d) Generating family F = ц + f*(u2 - t2fudu (k > 1), С = {(x, t) | (t2 - x2)x = 0}, az\Crtg = dF\Cl4 = (dr, + с ¦ xMdr2)|c^ for some с е С - {0}, ac(S2 - |cf* • SiJ = \c2tf ¦ lc. These are data as in Definition 4.23 for h{2k + 2), the map nc : С -*¦ M has degree 3, the map q : С -*¦ L covers one component with degree 1, the other with degree 2. Chapter 5 Singularities and Coxeter groups In this section several families of massive F-manifolds which come from sin- singularity theory are studied. The most important ones are the base spaces of semiuniversal unfoldings of hypersurface singularities. Three reasons for this are: A) hypersurface singularities and their unfoldings are so universal objects; B) their F-manifolds can be enriched to Frobenius manifolds (part 2); C) one has a 1-1 correspondence between irreducible germs of massive F-manifolds with smooth analytic spectrum and stable right equivalence classes of singular- singularities (Theorem 5.6). This is covered in section 5.1. The discussion of boundary singularities and their F-manifolds in section 5.2 is quite similar. Sections 5.3 and 5.4 are devoted to finite irreducible Coxeter groups and their F-manifolds and Frobenius manifolds. The discriminant in the complex orbit space induces an F-manifold structure on the orbit space just as in Corollary 4.6. This follows independently from work of Dubrovin and from results in singu- singularity theory by Brieskorn, Arnold, O.P. Shcherbak, Givental. We extend work of Givental in order to characterize these F-manifolds (Theorems 5.20, 5.21, 5.22) and use this to prove a conjecture of Dubrovin about the corresponding Frobenius manifolds (Theorem 5.26). In section 5.5 other families of F-manifolds with quite different properties are constructed. A start is made on the classification of 3-dimensional germs of massive F-manifolds. 5.1 Hypersurface singularities A distinguished class of germs of massive F-manifolds is related to isolated hypersurface singularities: the base space of a semiuniversal unfolding of an isolated hypersurface singularity is an irreducible germ of a massive F-manifold with smooth analytic spectrum (Theorem 5.3). In fact, there is a 1-1 61
62 Singularities and Coxeter groups correspondence between such germs of F-manifolds and singularities up to stable right equivalence (Theorem 5.6). The structure of an F-manifold on the base space has excellent geometric implications and interpretations (Theorem 5.4, Remarks 5.5). Many of these have been known for a long time from different points of view. The concept of an F-manifold unifies them. On the other hand, for much of the general treatment of F-manifolds in this book the singularity case has been the model case. An isolated hypersurface singularity is a holomorphic function germ / : (C"\ 0) -> (C, 0) with an isolated singularity at 0. Its Milnor number д 6 N is the dimension of the Jacobi algebra Oe",o/(ff-, ¦ • •, ^-) — Oc.o/Jf- The notion of an unfolding of an isolated hypersurface singularity goes back to Thorn and Mather. An unfolding of / is a holomorphic function germ F:(C"xC",0)^ (C, 0) such that F|c»,x(o) = /. The parameter space will be written as (M, 0) = (<C\ 0). The critical space (C, 0) С (Cm x M, 0) of the unfolding F - F(xu ... ,xm, ti,...,tn)is the critical space of the map (F, pr): (Cm хМ,0)ч- (С х М, 0). It is the zero set of the ideal ) 8хх'""дхя) E.1) with the complex structure Oc,o = Oc«xm,o/Jf\(C,0)- The intersection С П (<С'И х {0}) = {0} is a point and (C, 0) is a complete intersection of dimension n. Therefore the projection pr : (C, 0) -> (M, 0) is finite and flat with degree /j, and Oc,o is a free CVo-module of rank pi. A kind of Kodaira-Spencer map is the O^.o-linear map Oc,o, X ь* X(F)|(C,0) E.2) where X is any lift of X e TM,o to (Cm x M, 0). Dividing out the submodules шк,о • Тм,о and m^.o • Cc,o one obtains the reduced Kodaira-Spencer map ac|o:7bAf-> Oc~,o/Jf. E.3) All these objects are independent of the choice of coordinates. In fact, they even behave well with respect to morphisms of unfoldings. There are several possibilities to define morphisms of unfoldings (cf. Remark 5.2 (iv)). We need the following. Let Ft : (Cm x M-,, 0) -> (С, 0), * = 1, 2, be two unfoldings of / with projections prt : (<Cm x M,, 0) -»• (Af,-, 0), critical spaces C,, and Kodaira- Spencer maps ac, ¦ A morphism from F\ to F2 is a pair (ф, фьа*е) of map germs 5.7 Hypersurface singularities such that the following diagram commutes, (<Cra хМ,,0)Л(С'" хМ2,0) (Afb0) ^ (Af2,0), and 63 E.4) #lc«x{0} = id, F\ = F2 о ф E.5) E.6) hold. One says that F\ is inducedby (ф, фьа*е) from F2. The definition of critical spaces is compatible with the morphism (ф, фь^е), that is, фЧр2 = У/г, and (Сь 0) = ф~1((С2, 0)). Also the Kodaira-Spencer maps behave well: the CV,,o-hriear maps 0Ml,o Oci.o are defined in the obvious way; their composition is ac, = 0*I(C2,O) ° ac2 о йфЬсае. E-7) E.8) E.9) E.10) Formula E.9) restricts to the identity on the Jacobi algebra of / because of E.5). Therefore the reduced Kodaira-Spencer maps satisfy c, lo = ac2lo E.11) An unfolding of / is versal if any unfolding is induced from it by a suitable morphism. A versal unfolding F : (<Cm x M, 0) —> (C, 0) is semiuniversal if the dimension of the parameter space (M, 0) is minimal. Semiuniversal unfoldings of an isolated hypersurface singularity exist by the work of Thorn and Mather. Detailed proofs can nowadays be found at many places, e.g. [Was][AGVl, ch. 8]. Theorem 5.1 An unfolding F : (<Cm x M, 0) -> (C, 0) of an isolated hyper- hypersurface singularity f : (<Cm, 0) —> (C, 0) is versal if and only if the reduced Kodaira-Spencer map ado : TqM —> Oc»,o/Jf is surjective. It is semiuniver- semiuniversal if and only г/ас lo и ап isomorphism.
64 Singularities and Coxeter groups Remarks 5.2 (i) Because of the lemma of Nakayama ac lo is surjective (an isomorphism) if and only if ac is surjective (an isomorphism). (ii) A convenient choice of a semi universal unfolding F : (Cm xC,O) —> (C, 0) is F{x\,..., xm, ti,...,tn) = f + ?f=i ntitj, where mu ..., wM € Ос™,о represent a basis of the Jacobi algebra of /, preferably with ТП\ = 1. (iii) The critical space of an unfolding F : (Cm x C", 0) -> (C, 0) is reduced and smooth if and only if the matrix (-X?-, т^г)@) has maximal rank m. This OXj OXj OX'i Otfc is satisfied for versal unfoldings. (iv) In the literature (e.g. [Was]) one often finds a slightly different notion of morphisms of unfoldings: An (r)-morphism between unfoldings Fi and Fi as above is a triple (</>, <pbase, f) of map germs ф and фь^е with E-4) and E.5) and r : (Af i, 0) -> (C, 0) with E.6) replaced by = F2 о ф + т. E.12) The (r)-versal and (r)-semiuniversal unfoldings are defined analogously. They exist because of the following fact [Was]: an unfolding F : (Cm x M, 0) -> (C, 0) is (r)-versal ((T-)-semiuniversal) if and only if the map C®T0M ->Oc»,o/Jf, (c, X) y+с E.13) is surjective (an isomorphism). So one gains a bit: the base space of an (T-)-semiuniversal unfolding Fw has dimension д — 1; if F(r) = F(r\x\,..., xm, t2,..., fM) is (T-)-semiuniversal then f i + F^ is semiuniversal; between two semiunversal unfoldings t\ + F,(r) and t\ + Fj of this form there exist isomorphisms which come from (r)- isomorphismsof F,(r) and F2(r). (The relation between F(r) andti+F(r) motivates the '(>")'» which stands for 'restricted'.) On the other hand, one loses E.10). Anyway, one should keep (r)-semiuni- (r)-semiuniversal unfoldings in mind. They are closely related to miniversal Lagrange maps (see the proof of Theorem 5.6 and [AGV1, ch. 19]). (v) One can generalize the notion of a morphism between unfoldings if one weakens condition E.5): Let F, : (Cm x M,), 0) ->¦ (C, 0), i = 1, 2, be unfold- unfoldings of two isolated hypersurface singularities Д and /2. A generalized mor- morphism from Ft to F2 is a pair (ф, фь^е) of map germs with a commutative diagram as in E.4) such that E.6) holds and 0|c»x{oj is a coordinate change (between /1 and /2). Then /1 and /2 are right equivalent. If the generalized morphism is invertible then F\ and F2 are also called right equivalent. Critical spaces and Kodaira-Spencer maps also behave well for generalized morphisms; E.10) holds, in E.11) one has to take into account the isomorphism 5.7 Hypersurface singularities of the Jacobi algebras of /1 and f2 which is induced by ф\с°х[0)- 65 The multiplication on the base space of a semiuniversal unfolding was first defined by K. Saito [SK6, A.5)][SK9, A.3)]. Theorem 5.3 Let f : (Cm, 0) —> (C, 0) be an isolated hypersurface singularity and F : (Cm xM,0)-> (C, 0) be a semiuniversal unfolding. The Kodaira-Spencer map ас '¦ Тмл ~> C*c,o is an isomorphism and induces a multiplication on Tm,q- Then (M, 0) is an irreducible germ of a massive F- manifold with smooth analytic spectrum, and E := a^1 (F|c) is an Euler field of weight 1. Proof. The map ac : TMt0 -> 0c,o is an isomorphism because of Theorem 5.1 and Remark 5.2 (i). The critical space (C, 0) is reduced and smooth. One applies Corollary 4.21 to (Z, 0) = (Cm x M, 0) and az = dF. The map q : (C, 0) -> (L, 7t~l@)) is an isomorphism, and тг~'(О) is a point. Theorem 4.20 (c) shows that F\c о q-1 is a holomorphic generating function. Therefore ? is an Euler field of weight 1. ? Theorem 5.4 Let f : (Cm,0)—> (C, 0) be an isolated hypersurface singularity and Fi : (Cm x M,-, 0) -> (C, 0), i = 1,2, be two semiuniversal unfoldings. There exists a unique isomorphism <p : (Mi, 0) —> (M2, 0) of F-manifolds such that фьсие = <Pfor апУ isomorphism (ф, фьам) of the unfoldings F] and Fj. Proof. The map фь^е '¦ W\, 0) -> (M2,0) is an isomorphism of F-manifolds because of E.10). Suppose that Ft = F2 and (Mi, 0) = (M2, 0). The tangent map of фъаж on Г0М1 is dфbase\o = id because of E.11). The group of all automophisms of (Mi, 0) as F-manifold is finite (Theorem 4.14). Therefore фЬа1е — id. ? Remarks 5.5 (i) The rigidity of the base morphism фьа5е in Theorem 5.4 is in sharp contrast to the general situation for deformations of geometric objects. Usually only a part of the base space of a miniversal deformation is rigid with respect to automorphisms of the deformation. (ii) The reason for the rigidity is, via Theorem 4.14 and Theorem 3.19, the existence of the canonical coordinates at generic parameters. The corresponding result for singularities is that the critical values of F form coordinates on the base at generic parameters. This has been proved by Looijenga [Lol]. (iii) Because of this rigidity the openness of versality (e.g. [Te2]) also takes a special form: For any point t € M in a representative of the base space
66 Singularities and Coxeter groups (M, 0) = (CM, O)ofasemhmiversal unfolding F, Theorem 2.11 yields a unique decomposition (A/, t) = fliLi^b 0 into a product of irreducible germs of F-manifolds. These germs (M^, t) are the base spaces of semiuniversal unfoldings of the singularities of F\c»x{t]. The multigerm of F at Cm x {t} П С itself is isomorphic - in a way which can easily be made precise - to a transversal union of versal unfoldings of these singularities. (iv) The tangent space T,M = 0*=1 T,Mk is canonically isomorphic to the direct sum of the Jacobi algebras of singularities of F |e» * (/) • The vector in T, M of the Euler field E is mapped to the direct sum of the classes of the function Лс»х{/} in these Jacobi algebras. A result of Scherk [Sche2] says: The Jacobi algebra Oc,o/Jf of an isolated hypersurface singularity f : (Cm, 0) -> (C, 0) together with the class [/] e Oc-,o/J/ determines f up to right equivalence. This result shows that the base space M as an F-manifold with Euler field E de- determines for each parameter t e M the singularities of F|c»x{r) up to right equiv- equivalence and also the critical values. Theorem 5.6 will give an even stronger result, (v) The eigenvalues of Eo : T,M -> T,M are the critical values of F|c»x(r). Therefore the discriminant of the Euler field E is V = {t e M | (det(?o))(/) = 0} = яс(С Л F~\0)) E.14) and it coincides with the classical discriminant of the unfolding F. All the results of section 4.1 apply to this discriminant. Of course, many of them are classic in the singularity case. For example, Theorem 4.1 and the isomorphism q : С -> L from Corollary 4.21 yield an isomorphism between the development T> С РТ*М of the dis- discriminant and the smooth variety С П F~'@) which has been established by Teissier [Te2]. Implicitly it is also in [AGV1, ch. 19]. The elementary way in Corollary 4.6 in which the discriminant and the unit field determine the Jacobi algebras seems to be new. But the consequence from this and Scherk's result that the discriminant and the unit field determine the singularity (up to right equivalence) is known (compare below Theorem 5.6 and Remark 5.7 (iv)). Arnold studied the relation between singularities andLagrange maps [AGV1, ch. 19]. His results (cf. also [Hoe], [Phi, 4.7.4.1, pp. 299-301], [Ph2], [Wir, Corollary 10]) together with those of section 3.4 yield the following correspon- correspondence between unfoldings and certain germs of F-manifolds. Theorem 5.6 (a) Each irreducible germ of a massive F-manifold with smooth 5.1 Hypersurface singularities 67 analytic spectrum is the base space of a semiuniversal unfolding of an isolated hypersurface singularity. (b) Suppose, Ft : (C'"( x Af,-, 0) -> (C, 0), i = 1, 2, are semiuniversal un- unfoldings of singularities f : (C"\ 0) -> (C, 0) and (p:(Mu0)-> (M2, 0) is an isomorphism of the base spaces as F-manifolds. Suppose thatm\ < m2. Then a coordinate change f : (C', 0) -+ (C', 0) exists such that /,(*!,..., xmi) + x2mi+l + ..-+x2m2 = f2{xu ..., xm2) о ф E.15) and an isomorphism (ф, фьа*е) of the unfoldings Fy + x^i+1 -I h хгтг and Fivty exists with and E-16) Proof, (a) The restricted Lagrange map of the germ of a massive F-manifold with smooth analytic spectrum is a miniversal germ of a Lagrange map with smooth Lagrange variety (section 3.4). Arnold [AGV1, 19.3] constructed a generating family F(r) = Fir\x, t2,..., t^) for it. Looking at the notions of stable maps and generating families in [AGV1, ch. 19], one sees: F(r) is an (r)-semiuniversal unfolding of F(r){x, 0) (cf. Remark 5.2 (iv)). The unfolding tx 4- f M is a semiuniversal unfolding of F{r)(x, 0). Its base space is the given germ of a massive F-manifold. (b) The unfolding F,- is isomorphic to an unfolding Ц + F/r)(*i,. ..,xmi, t2,.. ¦, tft.) as in Remark 5.2 (iv) over the same base. Then F;<r) is an (r)- semiuniversal unfolding and a generating family for the restricted Lagrange map of the F-manifold (Af,-, 0). The isomorphism (p : (Af i, 0) -> (Af2, 0) induces an isomorphism of the re- restricted Lagrange maps. Then the main result in [AGV1, 19.4] establishes a notion of equivalence for F\r) and F^, stable 7?+-equivalence, which yields the desired equivalence in Theorem 5.6 (b) for F| and F2. ? Remarks5.7 (i) Two isolated hypersurface singularities/; : (C"",0) -> (C, 0) with mi < m 2 are stably right equivalent if a coordinate change \jr : (СШ2,0) -> (С2, 0) with E.15) exists. Furthermore they are right equivalent if m, = m2. The splitting lemma says: An isolated hypersurface singularity f : (Cm, 0) -*¦ (C, 0)withr :=m- rank(g^?-)@) is stably right equivalent to a singularity g : (C, 0) -> (C, 0) with rank(g|^)@) = 0; this singularity g is unique up to right equivalence.
68 Singularities and Coxeter groups (For the existence of g see e.g. [SI, D.2) Satz], the uniqueness of g up to right equivalence follows from Theorem 5.6 or from Scherk's result (Remark 5.5 (iv)).) (ii) Theorem 5.6 gives a 1-1 correspondence between isolated hypersurface singularities up to stable right equivalence and irreducible germs of massive F-manifolds with smooth analytic spectrum. But the liftability of an isomorphism cp : (M\, 0) -» (Л/2, 0) to unfoldings which is formulated in Theorem 5.6 (b) is stronger. The 1-1 correspondence it- itself already follows from Theorem 5.6 (a) and Scherk's result (Remark 5.5 (iv)). (Hi) The proof of Theorem 5.6 (a) is not very difficult. If (M, 0) is an irre- irreducible germ of a massive F-manifold with analytic spectrum (L, А.) С Т*М, then a sufficiently generic extension of a generating function on (L, A.) to a function on (T*M, A.) is already a semiuniversal unfolding over (M, 0). A ver- version different from [AGV1,19.3] of the precise construction is given by Pham [Phi, 4.7.4.1, pp. 291-301], following Hormander [Hoe]. (iv) Theorem 5.6 (b) follows also from [Ph2] (again following Hormander) and from [Wir, Corollary 10]. To apply Wirthmiiller's arguments one has to start with the discriminant V and the unit field. Pham [Phi] [Ph2] starts with the characteristic variety. That is the cone in T*M of the development 2? с ?Т*М of the discriminant. A semiuniversal unfolding F : (Cm x M, 0) -» (C, 0) yields data as in Corollary 4.21 for the germ (Л/, 0) of an F-manifold: (Z, 0) = (Cm x M, 0), az = dF. E.17) The semiuniversal unfolding F is also a generating family of (M, 0) as a germ of an F-manifold in the sense of Definition 4.23. The following observation says that these two special cases Corollary 4.21 and Definition 4.23 of the general construction of F-manifolds in Theorem 4.20 meet only in the case of unfoldings of isolated hypersurface singularities. Lemma 5.8 LetZ, M,nz,C,oiz,ac,andF : Z -> С satisfy all the properties in Corollary 4.21 and Definition 4.23. Then С is smooth. For any point p e M the multigerm F : (Z,C П 1 ' 1 z (?)) -*¦ ^ is isomorphic to a transversal product ofversal unfoldings of the singularities ofF\n^\p) (cf. Remark5.5 (Hi)). The irreducible germs (Mb p) of F-manifolds in the decomposition (M, p) = Ц1к=](Мк, р) are base spaces of semiuniversal unfoldings of the singularities o ^ ' (p). 5.2 Boundary singularities 69 Proof. The isomorphism ac : TM -*¦ (,nc)*Oc of Corollary 4.21 restricts at p e M to a componentwise isomorphism of algebras (Jacobi algebra of F\(nz' (p), z)). One applies Theorem 5.1. ? 5.2 Boundary singularities The last section showed that germs of F-manifolds with smooth analytic spec- spectrum correspond to isolated hypersurface singularities. The simplest nonsmooth germ of an analytic spectrum of dimension n is isomorphic to We will see that irreducible germs of massive F-manifolds with such an analytic spectrum correspond to boundary singularities (Theorem 5.14). Boundary sin- singularities had been introduced by Arnold [Ar2]. Because of the similarities to hypersurface singularities we will take things forward exactly as in section 5.1. We always consider a germ (Cm+1,0) with coordinates xo,...,xm together with the hyperplane H := {x 6 Cm+1 | x0 = 0} of the first coordinate. A boundary singularity (/, H) is a holomorphic function germ / : (Cm+1, 0) ->¦ (C, 0) such that / and f\H have isolated singularities at 0. It can be considered as an extension of the hypersurface singularities / and f\H- Its Jacobi algebra is Eл8) and its Milnor number д = [Sz, §2]) Я) := dimOc^.o/J/.H satisfies ([Ar2, §3], E.19) An unfolding of (/, Я) is simply a holomorphic function germ F : (Cm+1 x C, 0) -> (C, 0) such that F|Cm+1 x {0} = /, that is, an unfolding of/. Again we write the parameter space as (M, 0) = (C\ 0). But the critical space (C, 0) С (Cm+1 x M, 0) of F as unfolding of the boundary singularity (/, Я) is the zero set of the ideal dF 3F dF\ J E20)
70 Singularities and Coxeter groups with the complex structure Oc,o = 0o»+' xM,o/jf,h\(c,O) (cf. [Sz]). Forgetting the complex structure, (C, 0) is the union of the critical sets (C(l), 0) of F and (CB), 0) of F\hxm as unfoldings of hypersurface singularities. For the same reasons as in the hypersurface case the projection pr : (C, 0) —> (M, 0) is finite and flat with degree /u. and Oc 0 is a free CV^-module of rank д. The 1-form az := - E.21) on (Z, 0) := (C"+I x M, 0) gives rise to a kind of Kodaira-Spencer map ac : TMi0 -»¦ Oc.o, X м- az(#)l(c,o)- E.22) where X is any lift of X e TM,oto(Z, 0).Itinduces a reduced Kodaira-Spencer map ac|o : T0M -»• Oe»xfi/Jf,H. E.23) The ideal Jf, h and the maps ac and ac |o behave well with respect to morphisms of unfoldings, as we will see. A morphism between two unfoldings F\ and F2 as in section 5.1 of a bound- boundary singularity (/, H) is a pair (ф, фь^е) of a map germ with E.4)-E.6) and additionally E.24) ф(Н x МО С Я x M2. Then the first entry of ф takes the form xq • unit 6 Oz,o- Using this one can see with a bit more work than in the hypersurface case that the critical spaces behave well with respect to morphisms: ф*]Рг = JF, and (C,,0) = Ф~\(С2, 0)). E.25) Also the Kodaira-Spencer maps behave as well as in the hypersurface case. The OWb0-linear maps AфЬа8е, ac2, and ф*\(с2,0) are defined as in E.7)-E.9); again one finds and ас, = Ф*\(сг,о) о ас2 о йфЬа] ac,lo = E.26) E.27) Versal and semiuniversal unfoldings of boundary singularities are defined anal- analogously to the hypersurface case and they exist. 5.2 Boundary singularities 71 Theorem 5.9 [Ar2] An unfolding F : (Cm+1 x M, 0) -> (C, 0) of a boundary singularity (/, H), f : (C+1, 0) -> (C, 0), is versal if and only if the reduced Kodaira-Spencer map ac|o •' T0M ->• Ос™+1,о/^/,я is surjective. It is semi- universal if and only if ado is an isomorphism. Remarks 5.10 (i) The map ado is surjective (an isomorphism) if and only if ac is surjective (an isomorphism). (ii) The function F(x0,..., xm, t{,..., fM) = / + ??., m,-f,- is a semiuni- semiuniversal unfolding of the boundary singularity (/, H) if m \,..., mM 6 Oo"+I 0 represent a basis of Oc"+',o/^/,«- (iii) The critical space of an unfolding F : (Z,0) = (Cn+1 xM,0)-+ (C, 0) of a boundary singularity (/, H) is reduced and isomorphic to ([(x, у) е C2 I xy = 0}, 0) x (С", 0) if and only if f?,..., |?- represent a generating system of the vector space mz^AOto) + m| 0). This is equivalent to the non- degeneracy condition rank гапк bxabxj d2F Ьх;Ьх; 32F Эх0Э1к 32F @) = m + 1 E.28) (cf. [DD]). It is satisfied for versal unfoldings. (iv) As in Remark 5.2 (v) for hypersurface singularities, one can define generalized morphisms between unfoldings of right equivalent boundary sin- singularities. Again the critical spaces and Kodaira-Spencer maps behave well. Theorem 5.11 Let F : (Cm+1 x M, 0) -> (C, 0) be a semiuniversal unfolding of a boundary singularity (/, #). The Kodaira-Spencer map &c • ^м,о ~~*" Oc,o is an isomorphism and induces a multiplication on 7м,о- Then (M, 0) is an irreducible germ of a massive F- manifold with analytic spectrum isomorphic to ({(x, y) 6 C2 | xy = 0}, 0) x (О*,0). The field E := ac(F|c) is an Euler field of weight 1. Proof. Similar to the proof of Theorem 5.3. One wants to apply Corollary 4.21 and has to show that <xz\creg is exact. The critical space (C, 0) as a set is the union of the smooth zero sets (CA\ 0) of Jf and (CB), 0) of the ideal . The definition E.21) of az shows (*o, f^, ¦..., azl(c«),0) = dF|(C(o,0) for i = 1, 2. E.29) ( Therefore az\creg is exact and F\c о q~' is a holomorphic generating function of the analytic spectrum. ?
72 Singularities and Coxeter groups Theorem 5.12 Let Ft : (C+1 x M,, 0) -> (C, 0), i = 1,2, be two semiuni- versal unfoldings of a boundary singularity (/, H). There exists a unique isomorphism <p : (M\, 0) —> (M2, 0) of F-manifolds such that fybase — <Pforany isomorphism (ф, фьа*е) of the unfoldings F\ and F2. Proof: Similar to the proof of Theorem 5.4. D Remarks 5.13 (i) Let F : (C+1 x M, 0) -* (C, 0) be a semiuniversal unfold- unfolding of a boundary singularity (/, H) with critical space (C, 0) = (CA\ 0) U (CB), 0). For any t e M the points in Cm+1 x [t} П (C, 0) split into three classes: The hypersurface singularities of F\Cm+1 x [t) in CA) - CB), the hypersurface singularities of F\H x {t) in CB) — CA), and the boundary singularities of F|Cm+1 x {t} in СA)ЛС<2). The algebra Oc|Cm+l x {0} is the direct sum of their Jacobi algebras. The reduced Kodaira-Spencer map at t e M is an isomorphism from T, M to this algebra. Hence the multigerms of F at Cm+1 x [t] П CA) and of F\H x M at H x {t) П (CB) — CA)) together form a transversal union of versal unfoldings of these hypersurface and boundary singularities. The components {Mk, t) of the decomposition (M, t) = \^к=\Шк, t) into irreducible germs of F-manifolds are bases of semiuniversal unfoldings of the hypersurface and boundary singularities. (ii) The eigenvalues of Eo : T,M —y T,M are by definition of E the values of F on Cm+1 x {t} П С The discriminant of the Euler field is V = [t e M | (det(?o))(f) = 0} = 7TC(C П F~' @)). E.30) This is the union of the discriminants of F and F\H x Mas unfoldings of hypersurface singularities and it coincides with the classical discriminant of F as an unfolding of a boundary singularity [Ar2][Sz]. All the results of section 4.1 apply to this discriminant. Nguyen huu Due and Nguyen tien Dai studied the relation between boundary singularities and Lagrange maps [DD]. Their results together with section 3.4 yield the following correspondence between unfoldings of boundary singular- singularities and certain germs of F-manifolds. Theorem 5.14 Let(M, 0) be an irreducible germ of a massive F-manifold with analytic spectrum (L, X) isomorphic to ({U, y) € C2 | xy = 0}, 0) x (С, 0) and ordered components (LA), X) U (LB), k) = (L, X). 5.2 Boundary singularities 73 (a) There exists a semiuniversal unfolding F of a boundary singularity such that the base space is isomorphic to (M, 0) as F-manifold and the isomorphism q : (C, 0) -*¦ (L, A.) maps C(l) to L(i\ (b) Suppose, Fi : (C"+1 x M,-,0) -»¦ (C, 0), i = 1,2, are semiuniversal unfoldings of boundary singularities (/|, Я,) and <p : (Mt, 0) -> (M2, 0) is аи isomorphism of the base spaces as F-manifolds. Suppose thatm\ < m2. ' Then a coordinate change i/r : (СШ2+1, 0) -> (Cm2+1, 0) with i/((H2,0)) = (Hz, 0) exisfs suc/z that fx(x0,.. .,xm) +xl1+l + • • • +xl2 = /2(* isomorphism (ф, фъа5е) of the unfoldings F} of boundary singularities exists with 0,... ,xm2) о E.31) -I h jc2 * = and E.32) Froq/: (a) In [DD, Proposition 1] an unfolding F : (Cm+1 x M, 0) -> (C, 0) with nondegeneracy condition E.28) of a boundary singularity is constructed such that F is a generating family for LA) С T*M and F \ H x M is a generating family for LB). One can show that there are canonical maps C(l) —v L(l) which combine to an isomorphism q : С —у L with &c = q о a (as in Theorem 4.20). Then the Kodaira-Spencer map ac : Тм,о -у Ос,о is an isomorphism and F is a semiuniversal unfolding of a boundary singularity. (Implicitly this is also contained in [DD, Theoreme]). Because of ac = q о a its base is (M, 0) as F-manifold. (b) [DD, Proposition 3]. D Remarks 5.15 (i) Two boundary singularities fx : (C""+1,0) -> (C, 0) with mi < m2 are stably right equivalent if a coordinate change i/r as in Theorem 5.14 (b) exists. Furthermore they are right equivalent if m\ = m2. A splitting lemma for boundary singularities is formulated below in Lemma 5.16. (ii) Theorem 5.14 gives a 1-1 correspondence between boundary smgularities up to stable right equivalence and irreducible germs of massive F-manifolds with analytic spectrum (L, X) = {{(x, y) € C2 \xy - 0}, 0) x (C2,0) and ordered components (LA), k) U (LB), X) = (L, X). (iii) Interchanging the two components of (L, X) corresponds to a duality for boundary singularities which goes much further and has been studied by I. Shcherbak, A. Szpirglas [Sz][ShSl][ShS2], and others.
74 Singularities and Coxeter groups Lemma 5.16 (Splitting lemma for boundary singularities) A boundary singularity (/, H)with f : (C'"+1, 0) -+ (C, 0)andH = {x\xo =0} is stably right equivalent to a boundary singularity g : (C+1, 0) —*¦ (C, 0) in @) E.33) г + 1 = maxl 2; т + 1 - rank coordinates. The boundary singularity g is unique up to right equivalence. Proof. Existence of g: The group G = Z2 acts on (C!+1, 0) by (x0, xt,..., xm) \-± (±x0, X], ...,xm). Boundary singularities on the quotient (Cm+1, 0) correspond to G-invariant singularities on the double cover, branched along H ([AGV1, 17.4]). One applies an equivariant splitting lemma of Slodowy [SI, D.2) Satz] to the G-invariant singularity f(Xg, xt xm). The nondegenerate quadratic part of the G-invariant singularity in splitted form does not contain x\ because / is not smooth. Uniqueness of g: This follows with Theorem 5.14 (b). D The following two observations give some information on generating families in the sense of Definition 4.23 for the F-manifolds of boundary singularities. The first one gives a distinguished generating family and is essentially well known. The second one explains why B2 = /2D) is missing in Example 4.24 (b). Lemma 5.17 (a) Let F : (Z, 0) = (Cm+1 xM,0)-> (C, 0) be a semiuniversal unfolding of a boundary singularity (/, H). Then the function F : (Z, 0) = (Cm+I x M, 0) ->¦ (C, 0) with F(x, t) = F(x%, x\,...,xm,t) is ageneratingfamily for the germ (M, 0) of an F-manifold. The finite map q : С -у L = LA) U LB) from its critical set С to the analytic spectrum L has degree 2 on LA) and degree 1 on LB). The branched covering С ^ Mhas degree 2fi(f) + n(f\H)- (b) Let (M, 0) be a germ of a massive F-manifold with analytic spectrum (L, X) ~ ([(x, y) € C2 | xy = 0}, 0) x (С, 0). There does not exist a gen- generating family F : (Z, 0) —у (С, 0) with critical set С such that the canonical map q : С —у L is a homeomorphism. Proof: (a) Consider the branched covering лгс : Z -> Z, (xo,..., xm, t) н> (Xq, x\ ,..., xm, t) which is induced by the action (x0, ...,xm,t) н* (±*о. X\,..., xm, f) of the group G = Z2 on Z. The composition F = F о лс is an unfolding of the G-invariant singularity F|(Cm+I x{0}, 0), infact, semiuniversal 5.5 Coxeter groups and F-manifolds ithin the G-invariant unfoldings (cf. [SI, D.5)]). The ideals / /3F \ 3F 3F \ \ \dxQ ) dx{ dxm ) 75 E.34) nd tvqJf.h have the same zero sets С = UqX{C). Comparison of a^ — dFonZandaz = — |j-dxo+dFonZ(formulaE.21)) shows that the map a^ : Тм,o —*¦ Cc.o factorizes into the Kodaira-Spencer map shows th p ^ , . ас '- 1~м,о -> Oc,o and the map Jr?|(c,o) : Oc,o -* C»c,o- lilii irit с - 1м,о > Oc,o and the map Jr?|(c,o) c,o c,o Therefore ag is injective with multiplication invariant image and induces the correct multiplication on Тц,й- The rest is clear. (b) Assume that such a generating family F exists. The analytic spectrum (L, X) is its own maximalization. Therefore the homeomorphism q is an iso- isomorphism. Then ac = q ° a (cf. Definition 4.23 and Theorem 4.20) is an isomorphism. We are simultaneously in the special cases Definition 4.23 and Corollary 4.21 of Theorem 4.20. By Lemma 5.8 (L, X) is smooth, a contradiction. D 5.3 Coxeter groups and F-manifolds The complex orbit space of a finite irreducible Coxeter group is equipped with the discriminant, the image of the reflection hyperplanes, and with a certain dis- distinguished vector field (see below), which is unique up to a scalar. Together they induce as in Corollary 4.6 the structure of an F-manifold on the complex orbit space (Theorem 5.18). This follows independently from [Du2][Du3, Lecture 4] and from [Gi2, Theorem 14]. In fact, both give stronger results. Dubrovin established the structure of a Frobenius manifold. This will be discussed in section 5.4. Givental proved that these F-manifolds are distinguished by certain geometric conditions (Theorem 5.21). With one additional argument we will show that the germs of these F-manifolds and their products are the only germs of simple F-manifolds whose tangent spaces are Frobenius algebras (Theorem 5.20). This complements in a nice way the relation between Coxeter groups and simple hypersurface and boundary singularities. We will also present simple explicit formulas for these F-manifolds which are new for Я3 and H\ (Theorem 5.22). A finite Coxeter group is a finite group W of linear transformations of the Euclidean space W generated by reflections in hyperplanes. Each Coxeter group is the direct sum of irreducible Coxeter groups. Their classification and descrip- description can be found in [Co] or [Bou]. They are An (n > 1), Dn (n > 4), Ef,, ?7,
76 Singularities and Coxeter groups Et, Bn (n > 2), F4, G2, H3, Щ, I2(m) (m > 3) with A2 = /2C), B2 = /2D), H2 := /2E), G2 = /2F). The Coxeter group W acts on C" = R" ®R С and on C[X],..., xn], where x\,... ,х„ are the coordinates on C". The ring С [xi,..., xn]w of invariant poly- polynomials is generated by n algebraically independent homogeneous polynomials P\,..., Pn. Their degrees d; := deg P,- are unique (up to ordering). The quo- quotient C/ W is isomorphic to C" as an affine algebraic variety. The C*-action and the vector field ?,. xt j^ on the original С induce a C*-action and a vector field ^,ditjj^ on the orbit space C/ W = C. The image in the orbit space of the union of the reflection hyperplanes is the discriminant V of the Coxeter group. Suppose for a moment that W is irreducible. Then there is precisely one highest degree, which is called the Coxeter number h. The degrees can be ordered to satisfy di = h > d2 > ... > dn-i >dn=2, di+dn+1-i =h + 2. The vector field e := ^- is unique up to a scalar. E.35) E.36) Theorem 5.18 The complex orbit space M := C/W = С of a finite irre- irreducible Coxeter group W carries a unique structure of a massive F-manifold with the unit field e — ^- and the discriminant T>. The discriminant V corre- corresponds to the Eulerfield E.37) of weight 1. Proof. The uniqueness follows from Corollary 4.6. The existence follows from Dubrovin's result ([Du2][Du3, Lecture 4], cf. Theorem 5.23) or Givental's result [Gi2, Theorem 14] together with Theorem 3.16. Below in Theorem 5.22 we will follow Givental and reduce it to classical re- results on the appearance of discriminants in singularity theory ([Bril] [Ar2] [Ly2] [ShO]). D Remarks 5.19 (i) Corollary 4.6 gives probably the most elementary way in which e and V determine the multiplication on the complex orbit space M = C/ W = C, at least at a generic point: the e-orbit of a generic point p e M 5.3 Coxeter groups and F-manifolds 11 intersects V transversally in n points. One shifts the tangent spaces of V at these points with the flow of e to TpM. Then there exists a basis eit... ,е„ of TPM such that Ya=\ e< = e and such that the hyperplanes 0,-^ С ¦ e,-, ; = 1,..., n, are the shifted tangent spaces of V. The multiplication on TpM is given by ei oej = Sijej. (ii) The unit field e = ^ is only unique up to a scalar. The flow of the Euler field respects the discriminant V and maps the unit field e and the multiplication to multiples, because of Lie?(e) = —e and Lie?(o) = o. Therefore the isomorphism class of the F-manifold (M, о, е, Е) is indepen- independent of the choice of the scalar. (iii) The complex orbit space of a reducible Coxeter group W is isomorphic to the product of the complex orbit spaces of the irreducible subgroups. The discriminant decomposes as in Remark 4.2 (v). Now any sum of unit fields for the components yields a unit field for C/W. The choices are parameterized by (С*I'гг ™*«га"'. But the resulting F-manifold is unique up to isomorphism. It is the product of F-manifolds for the irreducible subgroups. This F-manifold and its germ at 0 will be denoted by the same combination of letters as the Coxeter group. Theorem 5.20 Let ((M, p), o, e) be a germ of a massive F-manifold. The germ ((M, p), o, e) is simple and TpM is a Frobenius algebra if and only if((M, p), o, e) is isomorphic to the germ at 0 of an F-manifold of a finite Coxeter group. This builds on the following result, which is a reformulation with section 3.4 of a theorem of Givental [Gi2, Theorem 14]. Theorems 5.20, 5.21, and 5.22 will be proved below in the opposite order. Some arguments on Я3 and Щ in the proof of Theorem 5.21 will only be outlined. Theorem 5.21 (Givental) (a) The F-manifold of a finite irreducible Coxeter group is simple. The analytic spectrum (L, X) of its germ at 0 is isomorphic to ({(x,y) eC2|jc2=: /},0) x (Си-\0) withr= lforAn, Dn, En, r = 2for Bn, F*. r = 3/or Я3, H4 and r = m - 2 for him). (b) An irreducible germ of a simple F-manifold with analytic spectrum iso- isomorphic to a product of germs of plane curves is isomorphic to the germ at 0 of an F-manifold of a finite irreducible Coxeter group. Finally, we want to present the F-manifolds of the finite irreducible Coxeter groups explicitly with data as in Corollary 4.21. We will use the notations of Corollary 4.21. The following is a consequence of results in [Bril][Ar2][Ly2]
78 Singularities and Coxeter groups [ShO] on the appearance of the discriminants of Coxeter groups in singularity theory. Theorem 5.22 (a) The germs at 0 of the F-manifolds of the Coxeter groups An, D,,, E,,, Bn, F4 are isomorphic to the base spaces of the semiuniversal unfoldings of the corresponding simple hypersurface singularities A,,, Dn, En and simple boundary singularities Bn (or Cn) and F4. (b) For the F-manifolds (M, о, е) = (С, 0, e) of the finite irreducible Cox- Coxeter groups, a space Z with projection itz : Z -*¦ M, a subspace С С Z and a 1-form az will be given such that the map X н* az(X)\c E.38) is welldefined and an isomorphism of Qм-algebras. The space С is isomor- isomorphic to the analytic spectrum of (M, o, e). The Euler field is always E = \ E"=i diUw,- The discriminant V С М is V = лс(ас(Е)~1 (О)). (i) А„, В„.'H3, him) : Z = CxM = CxC" with coordinates (x, t) = (x, t\,..., tn), az=dti l , t) := - \)x .n-1. h }, Bn:C = {(x,t)eZ\x- (xn~x - g) = 0}, H3: С = {(x, f)eZU3-? = 0}, him) :C = {(x,t)eZ\x2- 72m~2 = 0}. (it) D4, F4, H4 : Z = C2 x M = C2 x C4 with coordinates (x, y, t) = (x, y, t\,.. az — dt\ + xdt2 + ydt-i + xydt4, h(x, y, t) := h + ytt, b(x, У, t) ¦¦= h + xt4, D4:C = {(x,y,t)e Z\x2 + f2=0,y2 + h = 0}, F4:C = {(x,y,t)eZ\x2 + r2= 0, у2 + Г32 = 0}, H4:C = {{x, y, t)eZ\x2 + r2= 0, y2 + t? = 0}. (iii)Dn,E6,E7,E8: Z — C2 x M = C2 x С with coordinates (x, y, t) ;= (x, y, t\,.. F : Z -*¦ Си semiuniversal unfolding of F\C2 x {0}, Dn: F=xn-' E6: F = ., tn), = |f =0}, yh +x2t4 + xyts + x2yt6, 5.3 Coxeter groups and F-manifolds 79 F = хъу F = x5 + у3 + t\ / +1\ + xt2 + yh + x2t4 + xyt5 + x3f6 + x4t7, xt2 + yt3 + x2t4 + xyts + x3t6 + x2yt7 Proof of Theorem 5.22: (a) One can choose a semiuniversal unfolding F = f(xi,..., xm) + Yl"=] m'fi of the hypersurface or boundary singularity which is weighted homogeneous with positive degrees in all variables and parameters. There is an isomorphism from its base space C" to the complex orbit space of the corresponding Coxeter group which respects the discriminant, the Euler field, and the unit field ([Bril][Ar2]). It also respects the F-manifold structure (Corollary 4.6). (b) Part (i) for /2(m) is Remark 4.24 (a). Part (i) for А„, В„ and part (ii) for D4 follow with (a), with semiuniversal unfolding as in (a) for the singularities __i_x»+i (A,,), (_ij.» + yit н = {x = 0}) {В„), \хг + |v3 (D4). Also (iii) follows with (a). The same procedure gives for the boundary singularity F4 with equation (|y2 + |x3, H = {y = 0}) the data in (ii) with critical set 2 + T2 = 0, у2 С = {{x,y,t)€Z\x2 уГ3 = 0}. E.39) It is a nontrivial, but solvable, exercise to find compatible automorphisms of Z and M which map С to С andaz to «z modulo /c^z. Independently of explicit calculations, the proof of Theorem 5.21 will show that the data (Z,ctz, C) in (ii) correspond to F4. The data in (i) for Щ and in (ii) for H4 can be obtained from results of O.P. Shcherbak [ShO, pp. 162 and 163] (cf. also [Gi2, Proposition 12]) (for Я3 one could use instead [Ly2]). The unfoldings = f Jo xJdu+tl+xt2+x2t3 of ZN and p«4 = / (и2 Jo t3 +xt4fdu t\ + xt2 E.40) E.41) of Es have only critical points with even Milnor number and are maximal with this property. Their discriminants are isomorphic to the discriminants of the Coxeter groups H3 and H4. The unfoldings are generating families in the sense of Definition 4.23 for the F-manifolds of the Coxeter groups H3 and H4. We will determine the data in (ii) for H4 from F#4; the case of H3 is similar. Consider the map ф : C2 x C4 -> Z = C2 x С4, (х,у, t) ь» (х,у, t), E.42)
80 Singularities and Coxeter groups y(x, y, t) := / 2(«2 + f3 + xt4)du = -y3 + 2(f3 + хц)у, E.43) and observe ytA)dx + (y2 + f3 + At i + xdt2 + y&h + xydt4, 9 2 3 = (У2 Therefore V ф*(а2) = dF - ^dbc - ^-dy dx dy and the image under ф of the reduced critical set С г of FHi is E.44) E.45) E.46) = \(x, y, t) € Z|3*2 + t2 + yt4 = 0, —y2 + h +xt4 = 0 . 16 E.47) An automorphism s~x Z-> Z, (x, y, t\,t2, h, f4) (-»• {r~lx, s~xy, tu rti, st3, rst4) E.48) for suitable r, s € C* maps ф(Ср) to С and respects 7tz and az, together with the induced automorphism M -+ M. ? Sketch of the proof of Theorem 5.21: (a) Consider the data in Theorem 5.22 (b). The Euler field on M = С is E = { ?(. rf,f, ^. The coefficients of the Lyashko-Looijenga map Л : M —*¦ C" are up to a sign the symmetric polyno- polynomials in the eigenvalues of Ea. Because of Lie^Eo) = Eo, the coefficient Л,- is weighted homogeneous of degree i with respect to the weights (f-, • ¦ •. ^) for (ti,..., tn). The Lyashko-Looijenga map is a branched covering of degree E.49) and (M, o, e) is simple (Corollary 4.17 (b)). The analytic spectrum is isomorphic toC. (b) The dimension dim(A/t, p) of an irreducible germ in the decomposition (Л/, р) = П*=|(^*> Р) °f a germ °f a massive F-manifold is equal to the intersection multiplicity of T* M with the corresponding germ (L, A.*) of the analytic spectrum L. This number will be called the intersection multiplicity IM(Xk). 5.3 Coxeter groups and F-manifolds 81 (SM(g), q) denotes for any q e M the /x-constant stratum through g (section 4.4), and /(<?) the number of irreducible components of (M, q). For any sub- variety ScLwe have the estimates max(l(q)\q € яE)) < и + 1 - min(IM(a)\cr e 5), E.50) max(dim(SM(<7), ?) | ? 6 5) > dim 5, E.51) max(modM(M, q) \ q € S) + n + 1 - min(/M(<r) | ст € S) > dim5. E.52) Therefore, if M is simple then min(/M(<r) | a 6 S) < n + 1 - dim S E.53) for any subvariety 5 С L. Now suppose that ((Л/, р), о, e) is an irreducible germ ofa simple F-manifold and that ф : (M, p) ~+ П?=](^1> ^) is an 18отофЫ8т to a product of germs of plane curves (they are necessarily plane because of Proposition 3.7). If at least two curve germs were not smooth, e.g. (Cn_i, 0) and (С„, 0), then the intersection multiplicities IM(p) for points p in S\ :— Ф'ЧПы?^, О) х {0}) would be at least 4; but dim(Si, p) = n — 2, a contradiction to E.53). So, atmost one curve, e.g. (Cn, 0), is not smooth. The irreducible germs of F-manifolds which correspond to generic points of 7гEг) for S2 := 0~1(П!'=11(^1'0) x {0}) are at most 2-dimensional because of E.53). Therefore €<C2 \x2 = yr},0) E.54) for some r e N. If r > 4 and n > 3 then the set of possible intersection multi- multiplicities for points in S2 has a gap at 3 and a subvariety S3 С S2 exists with dim S3 = n - 2 and min(/A/(cr) | a € S3) > 4 [Gi2, p. 3266], a contradiction to E.53). Therefore r € {1,2, 3} or л < 2. If г e {1,2} then (A/, p) is the base space of a semiuniversal unfolding of a hypersurface singularity (r = 1, Theorem 5.7, [AGV1, ch. 19]) or boundary singularity (r = 2, Theorem 5.15, [DD]). Simplicity of their F-manifolds corre- corresponds to simplicity of the singularities. The simple hypersurface singularities are An, Dn, Еь, Ei, E$ [AGV1]. The simple boundary singularities are Bn, Cn, and F4 [Ar2] [AGV1]. The boundary singularities Bn and Cn are dual boundary singularities and have isomorphic discriminants and F-manifolds. The details of the case r = 3 [Gi2, pp. 3269-3271] are difficult and will not be given here. In that case the set of possible intersection multiplicities for points in S2 has a gap at 5. If n > 6 then a subvariety S4 С S2 exists with dimS4 = n - 4 and min(IM(a) \ a € S4) > 6, a contradiction to E.53). The
82 Singularities and Coxeter groups case r = 3 and n = 3 corresponds to #3, the case r = 3 and n — 4 corresponds to #4. D Proof of Theorem 5.20: It is sufficient to consider an irreducible germ {{M, p), 0, e). If it corresponds to a Coxeter group then it is simple (Theorem 5.21 (a)) and TpM is a Frobenius algebra (Theorem 5.22 (b)). Suppose that (M, p) is simple and that TpM is a Frobenius algebra. We will show by induction on the dimension n — dim M that the analytic spec- spectrum (L, A) is isomorphic to {{(x, y) e C2 \ x2 = /'}, 0) x (C"~', 0) for some reN. This is clear for n = 2. Suppose that n > 3. The maximal ideal of TpM is called m. The socle Апп^аДш) of the Gorenstein ring TpM has dimension 1, therefore Апптрм0^) 5 m and m2 ф 0. In the equations for the analytic spectrum (L, А) С 7^M one can eliminate fibre coordinates which correspond to m2 С ТрМ: the embedding dimension of (L, A) is m embdim(L, A) < л + dim —- < In — 2 E.55) (Lemma4.22). Then (L, A) = (C2,0) x (L", A") (Proposition 3.7). There exists A2 6 L close to A such that (L, A2) = (L, A) and^(A2)isnotinthee-orbitof p. Now for all q near p, but outside of the e-orbit of p, the germ (M, «jO is reducible because of modM(M, p) = 0. For all q near p the germ (M, <y) is simple and TqM is a Frobenius algebra (Lemma 2.2). One can apply the induction hypothesis to the irreducible component of (M, n{X.2)) which corresponds to A2. Its analytic spectrum (Z/, A') is isomorphic to a product of a smooth germ and a curve as above. Now (L, A) = (L, A2) == (C"-dimZ-', 0) x (I/, A'). One applies Theorem 5.21 (b). D 5.4 Coxeter groups and Frobenius manifolds K. Saito [SK3] introduced a flat metric on the complex orbit space of a finite irreducible Coxeter group. Dubrovin [Du2][Du3, Lecture 4] showed that this metric and the multiplication and the Euler field from Theorem 5.18 together yield the structure of a massive Frobenius manifold on the complex orbit space (Theorem 5.23). The Euler field has positive degrees. Dubrovin [Du2][Du3, p. 268] conjec- conjectured that these Frobenius manifolds and products of them are the only massive Frobenius manifolds with an Euler field with positive degrees. We will prove this conjecture (Theorem 5.25). Theorem 5.20, which builds on Givental's result (Theorem 5.21, [Gi2, Theorem 14]), will be crucial. 5.4 Coxeter groups and Frobenius manifolds 83 We use the same notations as in section 5.3. A metric on a complex manifold is a nondegenerate complex bilinear form on the tangent bundle. The flat standard metric on C" is invariant with respect to the Coxeter group W and induces a fiat metric g on M — V. Dubrovin proved the following with differential geometric tools [Du2][Du3, Lecture 4 and pp. 191 and 195]. Theorem 5.23 (Dubrovin) Let W be a finite irreducible Coxeter group with complex orbit space M = C/ W, Euler field E, a unit field e, and a multipli- multiplication о on M as in Theorem 5.18. The metric g on M —V with g(X, Y) := g(E о X, Y) E.56) for any (local) vector fields X and Y extends to aflat metric on M and coincides with K. Saito'sflat metric. (M, о, е, Е, g) is a Frobenius manifold. The Euler field satisfies UeE(g) = (l + !) g. E.57) There exists a basis of flat coordinates z\,. ¦ ¦, Zn on M with Zi@) = 0 and E.58) Remarks 5.24 (i) K. Saito (and also Dubrovin) introduced the flat metric g in a way different from formula E.56): The metrics g and g on M — V induce two isomorphisms T(M - V) -> T*(M - V). The metrics g and g are lifted with the respective isomorphisms to metrics g* and g* on the cotangent bundle T*(M - V). Then g* = 1ле,(Г) E.59) ([Du3, pp. 191 and 195]). (Here g* and g* are considered as @, 2)-tensors.) K. Saito introduced g with the formula E.59). (ii) Closely related to E.56) and E.59) is ([Du3, pp. 191 and 270]) J2IT1 • IT* = 8*(dQudQ2) = ifi(dQ, оdQ2). E.60) frf dXj dxt Here Qi, Q2 6 C[jci, ..., xn]w are W-invariant polynomials; dgi and dQ2 are inteфreted as sections in T*M; the multiplication о is lifted to T*M with the isomorphism TM —> T*M induced by g; is is the contraction of a 1-form withE.
84 Singularities and Coxeter groups The first equality is trivial. Equation E.60) is related to Arnold's convolution of invariants ([Ar3][Gil]). (iii) A Frobenius manifold as in Theorem 5.23 for a finite irreducible Coxeter group is not unique because the unit field and the multiplication are not unique. Contrary to the F-manifold, it is not even unique up to isomorphism. There is one complex parameter between (M, o, e) and (M, g) tobechosen: (M, о, e, Е, с -g) respectively (M, c-o, c~[ e, E, g) is a Frobenius manifold for any с е C*. (iv) We consider only Frobenius manifolds with an Euler field which is normalized by Lie? (o) = 1 ¦ о (compare Remark 2.17 (c)). The product ]~[ M,- of Frobenius manifolds (M,-, o,-, e,-, ?,, g,) also carries the structure of a Frobenius manifold if Lie^,. (g,-) = Dg, holds with the same number DeC for all i. This follows from Proposition 2.10, Theorem 2.15 and Remark 2.17 (c) (compare also[Du3,p. 136]). (v) Especially, the complex orbit space C/W of a reducible Coxeter group can be provided with the structure of a Frobenius manifold if the irreducible Coxeter subgroups have the same Coxeter number. The Frobenius manifold is not unique. The different choices are parame- parameterized by (С*)!'"™*»™"»! in the obvious way (cf. the Remarks 5.19 (iii) and 5.24 (iii)). The spectrum of a Frobenius manifold (M, o,e, E, g) is defined as follows (cf. Remark 9.2 e)). The Levi-Civita connection of the metric g is denoted by V. The operator V? :TM -*¦ Тм, Хн Vx?, acts on the space of flat fields ([Du3, p. 132], [Manl, p. 24]) and coincides there with -ad E. The set of its eigenvalues {w\,..., wn) is the spectrum ([Manl]). If —ad E acts semisimple on the space of flat fields then there exists locally a basis of flat coordinates Z\ zn with E = 9 iZi +n)— E.61) for some r, e C. The following was conjectured by Dubrovin ([Du2][Du3, p. 268]). Theorem 5.25 Let ((M, p), о, е, Е, g) be the germ of a Frobenius manifold with the following properties: generically semisimple multiplication; Lie?(o) =l-o and Lie?(g) = D ¦ g; * = ^z,- for a basis of flat coordinates Zi with Zi(p) = 0; positive spectrum (w\,... ,wn), that is, to,- > Ofor all i. E.62) 5.4 Coxeter groups and Frobenius manifolds 85 Then (M, p) decomposes uniquely into a product of germs at 0 of Frobenius manifolds for certain irreducible Coxeter groups. The Coxeter groups have all the same Coxeter number h = -pZ\- Proof. As in the proof of Theorem 5.21 (a), the hypotheses on the Euler field show that the Lyashko-Looijenga map Л : (Мь р) -»¦ С is finite and that the F-manifold (M, o, e) is simple. One applies Theorem 5.20 and Theorem 5.26. D Theorem 5.26 Let ((M, p), о, е, Е, g) be the germ of a Frobenius manifold such that ((M, p), o, e, E) is isomorphic to the germ at 0 of the F-manifold of a finite Coxeter group with the standard Euler field. Then the irreducible Coxeter subgroups have the same Coxeter number and ((M, p), o, e, E, g) is isomorphic to a product of germs at 0 of Frobenius man- manifolds for these Coxeter groups. Proof: First we fix notations. W is a finite Coxeter group which acts on V = C" and respects the standard bilinear form. The decomposition of W into I irre- irreducible Coxeter groups W\,..., Wi corresponds to an orthogonal decomposi- decomposition V = 0?=1 Vt. The choice of n algebraically independent homogeneous polynomials Pi,...,Pn € C[jci ,..., xn]w identifies the quotient M = V/ W with C". The quotient map xj/ : V -*¦ M decomposes into a product of quotient maps yjfk : Vk -»• Vk/ Wk = Mk. The F-manifold M = Y\'k=i Mk is the product of the F-manifolds Mk. Setting e := ?"_, xt -^ on У and ek := e| vt, the standard Euler field Ek on Mk is Ek — ^-Airk{ek). Here hk is the Coxeter number of Wk. The Euler field on M is ^=Е?'=Ею'^- E-63) {wi,..., wn} is the union of the invariant degrees of Wk, divided by hk. Now suppose that g is a flat metric on the germ (M, 0) such that ((M, 0), o, e, E, g)isagerm of aFrobenius manifold withLie?(g) = Dg. Consider a system of flat coordinates zi,..., Zn of (M, 0) with z,@) = 0. The space 0"=1 С ¦ ^ of flat fields is invariant with respect to adE ([Du3, p. 132], [Manl, p. 24]) and the space of affine linear functions С • 1 ф 0"=1 С • z; С Ом,о is invariant with respect to E. Because E vanishes at 0 even the subspace 0"=1 С • z,- is invariant with respect to E. The weights wi,..., wn of E are positive. Therefore the coordinates zi, • • •. zn can be chosen to be weighted homogeneous polynomials in C[f i,..., tn] of
86 Singularities and Coxeter groups degree wu ...,wn. Thus the spectrum of the Frobenius manifold is {ш| w,,}. It is symmetric with respect to ~, because of Lie?(g) = D ¦ g; hence 1 + y. = ? for all fc = 1,...,/. The Coxeter numbers are all equal, h :=h\ = •••=/!;, the Euler field E is E = j^dfie). E.64) It remains to show that g is induced as in Theorem 5.23 from a metric on V which it the orthogonal sum of multiples of the standard metrics on the subspaces V*. The operatorU = Eo : TM -» TM is invertible onM-P. The metric g on M-V with #(Х,У):=*(&Г1Ш,1Г) E.65) is flat ([Du3, pp. 191 and 194], [Manl]). It lifts to aflat metric g on V-f~\V). We claim that g extends to a flat metric on the union f~{(V) of the reflection hyperplanes. It is sufficient to consider a generic point p in one reflection hyperplane. Then the ^-orbit of V(p) intersects V in n points; there exist canonical coordinates Mi,..., un in a neighbourhood of \//(p) with e,- о ej = S^ei, g(e,, ej) = 0 for л E = usei + ^(m,- + r,)e,- for some r,- G C*. i=2 (Р,^р)) = ({и|щ=0},0). The map germ -ф : (У, р) —> (M, i/(p)) is a twofold covering, branched along (V, f(p)), and is given by (Й1,..., un) н> (гГ,2, и2,...,ип) = (мь ..., и„) for some suitable local coordinates m~i ,..., un on (V, p). Then !1 зй[)~8 о e,, e,) = ,, e,), ~/ Э 1 M,+r,ev ~/ Э 9 \ т^г.^г =0 for» for/>2, So g extends to a nondegenerate (and then flat) metric on V. The Coxeter group W acts as a group of isometries with respect to g. It remains to show that the vector space structure on V which is induced by ~g (and 0 € V) coincides with the original vector space structure. Then g is an orthogonal sum of multiples of the standard metrics on the subspaces V*, 5.5 3-dimensional and other F-manifolds 87 because each W-invariant quadratic form is a sum of Wk -invariant quadratic forms on the subspaces V* and they are unique up to scalars. Let?be the vector field on V which corresponds to the C*-action of the vector space structure induced by g. Then Li&gg = 2 ¦ g. Because of 1лъЕ(Ы) = U, LieE(g) = (l + f)g,and? = jdi/f(s) we also have Lieeg = 2 ¦ g for the vector field ?, which corresponds to the C*-action of the old vector space structure. The differences—? satisfies Lies-j(g) = 0 and is a generator of a 1-parameter group of isometries. As it is also tangent to the union of reflection hyperplanes, it vanishes. The vector field e = "e determines a unique space of linear functions on V and a unique vector space structure. D 5.5 3-dimensional and other F-manifolds The F-manifolds in sections 5.1-5.3 were special in several aspects: the analytic spectrum was weighted homogeneous and a complete intersection. Therefore an Euler field of weight 1 always existed, and the tangent spaces were Frobenius algebras. Furthermore, the stratum of points with irreducible germs of dimen- dimension > 3 had codimension 2. Here we want to present examples with different properties. A partial classi- classification of 3-dimensional germs of massive F-manifolds will show that already in dimension 3 most germs are not simple and do not even have an Euler field of weight 1. Examples of germs (M, p) of simple F-manifolds such that TpM is not a Frobenius algebra will complement Theorem 5.20. First, a construction which is behind the formulas for An, Bn, #3, /2(m) in Theorem 5.22 (b)(i) provides many other examples. Proposition 5.27 Fix the following data: (M, 0) = (C, 0), (Z, 0) = (C, 0) x (M, 0) with coordinates (x, t) = (x, h,..., tn\ the projection nz : (Z, 0) -> (M, 0), the 1-form az := dti + xdt2 H h xn~ldtn on Z, .; ;, the function 72(x, t) := t2 + 2xt3 H \-(n - l)xn~\, an isolated plane curve singularity (or a smooth germ) f : (C2, 0) with f(x, 0) = xn ¦ unit g C{x}, the subvariety С := {(x, t) e Z \ f(x, T2) = 0} С Z. (a) The map (C, 0) ac : ¦ Oc,o, az(X)\c E.66) (X is a lift of X to Z) is welldefined and an isomorphism of О'м ^-modules. The germ (M, 0) with the induced multiplication on Tuss is an irreducible germ of a massive F-manifold. Its analytic spectrum is isomorphic to
88 Singularities and Coxeter groups (С, 0) = (С"-1, 0) х (/"'(О), 0). For each t e M the tangent space T,M is is- omorphic to a product of algebras C{x}/(xk) and is a Frobenius algebra. (b)An Euler field of weight 1 exists on (M, p) if and only if the curve singu- singularity f(x, y) is weighted homogeneous. (c) Suppose that mult/ = n. Then the caustic is К = {t e M | t2 = 0}. The germ (M, t) is irreducible for all t 6 1С, so the caustic is equal to the ^.-constant stratum of(M, 0). The modality is mod^M, 0) = n -1 (the maximal possible). Proof, (a) The 1-form ctz is exact on Creg because of daz = dxd72. One can apply Corollary 4.21. (b) Corollary 3.8 (b). (c) For t2 = 0 fixed we have f(x, T2) = x" ¦ unit e C{x, t3,..., tn]. Thus the projection nc : С -*¦ M is a branched covering of degree n, with ж^' ({t \ t2 = 0}) = {0} x {t | t2 = 0} and unbranched outside of {t \ t2 = 0). The analytic spectrum is isomorphic to C. ? Remarks 5.28 (i) The function T2 is part of a coordinate system on T*M for a different Lagrange fibration: the coordinates У\ - У\, Уг - Уг, ft - У! ~ y'2~l for i > 3, E.67) • 1- (и — l)v?~2fn, Ti = t; for J > 3 satisfy yidti = da. E.68) The analytic spectrum of an F-manifold as in Proposition 5.27 is L = {(y, t)eT'M\yi = l, f(y2, Г2) = 0,у!=0 for i > 3}. E.69) It is a product of Lagrange curves. (ii) Another different Lagrange fibration is behind the formulas for D4, F4, #4 in Theorem 5.22 (b)(ii). There are many possibilities to generalize the construc- construction of the above examples. In dimension 3, there exist up to isomorphism only two irreducible commu- commutative and associative algebras, й(|) := С{х)/(хг) and E.70) Qa):=C{x,y}/(x2,xy,y2y, E.71) and 6(l) is a Frobenius algebra, gB) not. 5.5 3-dimensional and other F-manifolds 89 Theorem 5.29 Let (M, p) be an irreducible germ of a 3-dimensional massive F-manifold with analytic spectrum (L,X) С Т*M. (a) Suppose TpM = g(I). Then (L, X) has embedding dimension 3 or 4 and (L, X) = (C2, 0)x(C", 0) for a plane curve (C, 0) С (С2, 0)vwY/imult(C", 0) < 3. An Eulerfield of weight 1 exists if and only if(C, 0) is weighted homogeneous. (b)Suppose TpM = Qmand(L, X) = (C2, 0)x(C, 0)wf/zmult(C, 0) < 3. Then ((M, p), o, e) is one of the germs A3, Вт,, Щ. (с)Suppose TpM = Q{v>and(L,X) = (C2, 0)x(C", 0)vwf/imult(C", 0) = 3. Then the caustic 1С is a smooth surface and coincides with the ц,-соп- stant stratum; that means, TqM = 2A) for each q 6 1С. The modality is mod^(M, p) = 1 (the maximal possible). (d) Suppose TpM = QB). Then (L, X) has embedding dimension 5 and (L, X) = (C, 0) x (L(r), 0). Here (L(r), 0) is a Lagrange surface with embedding dimension 4. Its ring Oz,w,o is a Cohen—Macaulay ring, but not a Gorenstein ring. Proof, (a) One chooses coordinates (fb t2, Lemma 4.22) with e = -щ and 9f2 C- — = 1 for (M, p) (as in the proof of m С TpM, E.72) m2 с ТРМ. E.73) The dual coordinates on (T*M, T*M) are y\,...,yn. There exist functions e0, au a2, b0, b\, b2 e C{t2, f3} with \yi = = b2y\ y2 =a2y\ E.74) The Hamilton fields of the smooth functions y\ — \ and уъ - Yll=o Ъ\У2 are щ and|-+.. .in T*M. They are tangent to L. Therefore (L, X) = (C2, 0)x(C, 0) with (C", 0) = (L, X) П T*M (cf. Proposition 3.7). The statement on the Euler field is contained in Corollary 3.8 (b). (b) We have mult(C", 0) < 2; and the intersection multiplicity of (C, 0) with a suitable smooth curve is 3. So, (C, 0) is either smooth or a double point or a cusp. In the first two cases, one can apply the correspondence between F-manifolds and hypersurface or boundary singularities (Theorem 5.6 and Theorem 5.14) and the fact that A3, Biy and C3 are the only hypersurface or boundary singu- singularities with Milnor number 3.
90 Singularities and Coxeter groups Suppose (С, 0) is a cusp and ((M, p), o, e) is not #3. Then it is not simple becauseofTheorem5.20.Theц.-constantstratumS^ = {q e M \ TqM = Qm) is more than the e-orbit of p. It can only be the image in M of the surface of cusp points of L, because at other points q close to p the germ (M, q) is A\ or A\ x A2. So at each cusp point X' of L the intersection multiplicity of T*(y)M and (L, X') is 3. This property is not preserved by small changes of the Lagrange fibration (e.g. as in Remark 5.28 (i)). But Givental proved that a versal Lagrange map is stable with respect to small changes of the Lagrange fibration [Gi2, p. 3251, Theorem 3 and its proof]. This together with section 3.4 yields a contradiction. (c) In this case, at each point A/ (close to X) of the surface of singular points of L the intersection multiplicity of T*^M and (L, X') is 3 and the map germ я : (L, A.') —*¦ (M, n(X')) is a branched covering of degree 3. This implies all the statements. (d) If the embedding dimension of (L, X) is < 4 then (L, X) = (C2, 0) x (C, 0) for some plane curve (C, 0) by Proposition 3.7. Then(L, A.) is a complete intersection and the tangent spaces TPM are Frobenius algebras. So if TPM = <2B) then the embedding dimension of (L, X) is 5. The ring Ow),o is a Cohen-Macaulay ring because the projection L(r) —*¦ M(r) of the re- restricted Lagrange map is finite and flat. It is not a Gorenstein ring because TpM is not a Gorenstein ring. ? The next result provides a complete classification and normal forms for those irreducible germs (M, p) of 3-dimensional massive F-manifolds which satisfy TpM = gA) and whose analytic spectrum consists of 3 components. Part (a) gives an explicit construction of all those F-manifolds. Theorem 5.30 (a) Choose two discrete parameters p2, рз e Nwithp2 > рз > 2 and choose рз — l holomorphic parameters (go, g\,..., gPi-2) G С* х C3~2 with go ф 1 if pi = рз- Define (M, 0) := (C3, 0), (Z, 0) := (C, 0) x (M, 0) with coordinates (x,tu t2, t3) = (x, t), fx:=Ch:=t^,2f3:=k-S, С ¦¦= ULi c<- := UL K*. 0 e Z I x = f|} с Z, b2(t2, t3) := (ръ-g+h- f^) • (рз • g + h ¦ |^ -p2 ¦ t^2~n)-1 (b2 is a unit in C{t2, t3\), b[(t2, ?з) .'= —b2 ¦ p2 • ^2 > the 1-form az '¦= dfi + xdt2 + (b2x2 + b\x)dt3 on Z. 5.5 3-dimensional and other F-manifolds 91 (i) Then - for i = 1,2,3, E.75) «zlc, = d(f, + f,)\c, for i = 1, 2, 3. E.76) &c '¦ 1~м,о —> Octo, X (->¦ &z(X)\c E.77) Chj The map (X is a lift ofX to Z) is welldefined and an isomorphism ofOMfi-modules. The germ (M, 0) with the induced multiplication on 7м,о is an irreducible germ of a massive F-manifold with TqM =. <2A). Its analytic spectrum is isomorphic to (C,0) = (C2,0) x(C',0), with (C, 0) = (C, 0) П ({(*, t)\ti=t3= 0}, 0). E.78) (Hi) The caustic is K. = {t e M \ t2 = 0} and coincides with the bifurcation diagram В and with the ^-constant stratum. (iv) The functions t\ + fi\cr i = 1> 2, 3, combine to a function F : С —> С which is continuous on С and holomorphic on Creg = С П {(x, t) \t2 ф 0}. The Eulerfield E on M — fC with лс\м-к.(Е) = Лм-к is -t2^)~. E.79) dtj dt The following conditions are equivalent: (a) The function F is holomorphic on С and E is holomorphic on M, (P) one has рз = 2 or (p2 = рз > 3 and gt = Ofor 1 < i < рз — 2), (у) the curve (C, 0) is weighted homogeneous. (b) Each irreducible germ (M, p) of a massive F-manifold such that TPM = 2A> and such that(L, X) has 3 components is isomorphic to a finite number of normal forms as in (a). The numbers p2 and рз are determined by (L, X). The number of isomorphic normal forms is < 2p2 ifp2 > рз and < 6p2 ifp2 = рз. Proof, (a) (i) Direct calculation. (ii) The map ас is an isomorphism because b2 is a unit in €,{t2, ?з}. One can apply Corollary 4.21 because of E.76). For the analytic spectrum see Theorem 5.29 (a). (iii) The branched covering (C, 0) -> (M, 0) is branched along {(x, t)\x = t2 = 0}. Compare Theorem 5.29 (c). The generating function F : С -*¦ С has three different values on n^\t) forteM with t2 ф 0 because of E.76) and the definition of _/). Therefore K. = B.
92 Singularities and Coxeter groups (iv) Formula E.79) can be checked by calculation. The equivalence (a) (P) follows. Corollary 3.8 (b) shows (a) «=*¦ (y) (one can also see (P) «=> (y) directly). (b) We start with coordinates (r,, t2, t3) for (M, p) as in the proof of Theorem 5.29 (a). The proofs of Theorem 5.29 (a) and Lemma 4.22 (b) give a unique construction of data (Z, 0) = (C, 0) x (C3, 0), (C, 0) С (Z, 0) and az = dfi + xdt2 + (b2(t2, b)x2 + b{{t2, t3)x + bo(t2, t3))dt3 E.80) as in Corollary 4.21 for the germ (M, p) = (C3,0) of an F-manifold. The set (C,0) = U,3=i(C,>0) is the union of 3 smooth varieties, which project isomorphically to (C3,0), and is isomorphic to the product of (C2, 0) and (C, 0) П ({(jc, t)\ti=t3= 0}, 0). The components (C,, 0) can be numbered such that the intersection numbers of the curves (Q, 0) П ({(x, t) \ r, = t3 = 0}, 0) are p2 - 1 for i = 1, 2 and p3 — 1 for i = 1,3 and for i = 2,3. The numbers p2 and p3 are defined hereby and satisfy pi > p3 > 2. The 1-formaz is exact on Creg and can be integrated to a continuous function F : (C, 0) -> (C, 0) with F\c, = h + ft for a unique function f; e C{t2, t3]. Then Q = {(*, 01 x = |f} and Э/, Э/ E.81) We will refine (f\,t2, h) in several steps and change Z, C, az, /b /2, /3 ac- accordingly, without explicit mentioning. 1st step: The coordinates (fj, t2, t3) can be chosen such that (C, 0) -»• (C3, 0) is branched precisely over {f б С3 112 = 0}. 2nd step: The coordinate t\ can be changed such that /i = 0. Then Ci = {{x, t) | л: = 0} and С, П Q = |(х, 0 I x = 0 = M J = {{x, t)\x=Q,t2 = 0). E.82) Because of /i|c,nc,_= /ilc,nc, = 0, the functions /2 and f3 can be written uniquely as /2 = ff • g, f3 = tf • g with ^2, p3 > 1, g, g e C{f2, t3] - t2 ¦ C{t2, t3]. Now E.82) shows Ц = tf'~l ¦ unit and p-t > 2. Therefore pt = pi and g and g are units, with g@) # g@) if p2 = p3. 3rd step: The coordinate fe can be chosen such that f2 = f^. E.81) yields b0 = 0 for i = 1 and fc, = -b2 ¦ p2 ¦ ff ~' for i = 2 and 1г 47 —' 5.5 3-dimensional and other F-manifolds 93 for (' = 3. The first, third and fifth factor on the right are units, therefore we have ft = r = rf~2 unit. 4th step: The coordinate t3 can be changed such that g = YlfLo 8i • 4 + We have brought the germ (M, p) to a normal form as in (a). The numbering of C\, C2, C3 was unique up to permutation of C\ and C2 if p2 > p3 and arbi- arbitrary if p2 = p3. The choice of t2 was unique up to a unit root of order p2. Everything else was unique. ? Remark 5.31 Certain results of Givental motivate some expectations on the moduli of germs of F-manifolds, which are satisfied in the case of Theorem 5.30. An irreducible germ (M, p) of a massive F-manifold is determined by its restricted Lagrange map (L(r), A,(r)) -+ (Г*М(г), T*M^) -+ (M<r), /A>) (section 3.4). Suppose that (M, p) is 3-dimensional with TpM = Qm. Then (L^r\ Я(г)) decomposes into a product of two Lagrange curves, a smooth one and a plane curve (C, 0) (Theorem 5.29 (a), Proposition 3.7). If we fix only the topological type of the curve (C, 0), we can divide the moduli for the possible germs (M, p) into three pieces: (i) moduli for the complex structure of the germ (C, 0), (ii) moduli for the Lagrange structure of (C, 0), (iii) moduli for the Lagrange fibration in the restricted Lagrange map. Within the д-constant stratum SM = [q e M \TqM = QA)} of a representative M, the moduli of types (i) and (ii) are not visible because the Lagrange structure of the curve (C, 0) is constant along SM. But the moduli for the Lagrange fibration are precisely reflected by 5M because of a result of Givental [Gi2, proof of Theorem 3]: as a miniversal Lagrange map, the restricted Lagrange map is stable with respect to small changes of the Lagrange fibration which preserve the symplectic structure; that means, the germ of the Lagrange map after such a small change is the restricted Lagrange map of (M, q) for a point q e SM close to p. In view of Theorem 4.15 and Theorem 5.29 (b) and (c) there is one module of type (iii) if mult(C", 0) = 3 and no module of type (iii) if mult (C, 0) = 1 or 2. Fixing the complex structure of the plane curve (C, 0), the choice of a Lagrange structure is equivalent to the choice of a volume form. Equivalence classes of it are locally parameterized by Яд/1)((С, 0)) ([Gi2, Theorem 1], [Va5]), so the number of moduli of type (ii) is /л, — r (Theorem 3.5 (b)). It is
94 Singularities and Coxeter groups equal to the number of moduli of right equivalence classes of function germs / : (C2, 0) -* (C, 0) with (/"'@), 0) = (C, 0). The /a-constant stratum of a plane curve singularity in the semiuniversal un- unfolding is smooth by a result of Wahl ([Wah], cf. also [Matt]) and its dimension depends only on the topological type of the curve. So one may expect that the number of moduli of types (i) and (ii) together depends only on the topological type of (C, 0) and is equal to this dimension. (But a canonical relation between the choice of a Lagrange structure and the choice of a function germ for a plane curve (C, 0) is not known.) In the case of Theorem 5.30 these expectations are met: the topological type of the plane curve is given by the intersection numbers p2 — 1 and pi — 1; the last one gP3-2 of the complex moduli is of type (iii), it is the module for the ^-constant stratum and for the Lagrange fibration; the other рз — 2 moduli (go, • • •. 8п-з) are °f types (i) and (ii). One can check with [Matt, 4.2.1] that p2 — 2 is the dimension of the /u.-constant stratum for such a plane curve singularity. Finally, at least a few examples of germs (M, p) of F-manifolds with TqM = <2B> will be presented. Proposition 5.32 Consider M = C3 with coordinates (t\, t2, b) and T*M with fibre coordinates у i, y2, уз- Choose p2, Рз € N>2. Then the variety L = {(y, t) e T*M | yi = 1, У2(у2 - p2t?-]) = y2y3 1) = 0} E.84) has three smooth components and is the analytic spectrum of the structure of a simple F-manifold on M with TqM = gB). The field Э 1 3 1 Э E.85) is an Euler field of weight 1. Proof. One checks easily that a = y\dt\ + y2dt2 + уз&з is exact on the three components of L, that the map &c '¦ 1м —* л+Ol is an isomorphism of Од/- modules, and that E in E.85) is an Euler field. The weights of E are positive. This shows via the Lyashko-Looijenga map that (M, o, e) is a simple F-manifold (cf. the proof of Theorem 5.21 (a)). D Remark 5.33 In [Gi2, Theorem 15] the restricted Lagrange maps of two other series of simple F-manifolds with M = С and TqM not a Frobenius manifold 5.5 3-dimensional and other F-manifolds 95 are given, they are the series an (n > 3) and Я„ (и > 4) (also S| = Ab 32 = Hi, п2 = A2, ft3 = tf3). They have Euler fields of weight 1 with positive weights. The analytic spectra of 3,, and Q.n are isomorphic to С х ?„_|Bп - 1) and С2 х E,,_2B« - 3), respectively. Here E^B^+1) is the open swallowtail, the subset of polynomials in the set of polynomials [z2k+l + a2z2k~^ Л h a2k+\ \ a2,..., a2k+} s C} which have a root of multiplicity > к + 1 ([Gi2, p. 3256]). It has embedding dimension 2k. The germs ((M, 0), o, e) are irreducible for 3„ and Я„, the socle АппГоМ(щ) of TqM is the maximal ideal m С ТцМ itselfin the case of 3„ and has dimension n — 2 in the case of Я„. Givental [Gi2, Theorem 15] proved that the germs (M, 0) for 3„ and Я„ are the only irreducible germs of simple F-manifolds whose analytic spectra are products of smooth germs and open swallowtails. Generating functions in the sense of Definition 4.23 are due to O.P. Shcherbak and are given in [Gi2, Proposition 12].
Part 2 Frobenius manifolds, GauB-Manin connections, and moduli spaces for hypersurface singularities
Chapter 6 Introduction to part 2 The notion of a Frobenius manifold was introduced by Dubrovin in 1991 [Dul], motivated by topological field theory. It has been studied since then by him, Manin, Kontsevich, and many others. It plays a role in quantum cohomology [Man2] and in mirror symmetry. But the first big class of Frobenius manifolds had already been constructed in 1983 in singularity theory. K. Saito [SK6][SK9] studied the semiuniversal unfolding of an isolated hypersurface singularity and its GauB-Manin connec- connection. He was interested in period maps and defined the primitive forms as volume forms with very special properties in relation to the GauB-Manin connection. Any primitive form provides the base space of a semiuniversal unfolding of a singularity with the structure of a Frobenius manifold. He proved the existence of primitive forms in special cases and M. Saito proved their existence in the general case [SM2][SM3]. Using the work of Malgrange [Mal3][Mal5] on deformations of microdifferential systems, M. Saito showed that the choice of a certain filtration on the cohomology of the Milnor fibre yields a primitive form and thus a Frobenius manifold. This construction of Frobenius manifolds in singularity theory has been quite inaccessible to nonspecialists, because the GauB-Manin systems are treated using the natural, though sophisticated language of algebraic analysis and especially Malgrange's results require microdifferential systems and cer- certain Fourier-Laplace transforms. This also made it difficult to apply the construction. The first purpose of part 2 of this book is to give a detailed account of a sim- simplified version of the construction. This version stays largely within the frame- framework of meromorphic connections and is sufficiently explicit to work with it. The second purpose is to present several applications. The most difficult one is the construction of global moduli spaces for singularities in one д-homotopy class as an analytic geometric quotient. 99
100 Introduction to part 2 Outlines of the construction and the applications are offered in the following two sections. In these outlines the reader can jump straightaway to the main points. An orientation is given how all the material in the subsequent chapters is used and a motivation to study it. The applications are given in chapters 12, 13, and 14. In chapter 12 we give a canonical complex structure on the /г-constant stratum of a singularity and an infinitesimal Torelli type result, which strengthens a result of M. Saito. In chapter 13 the global moduli spaces for singularities are constructed and symmetries of singularities are discussed, extending some work of Slodowy and Wall. In chapter 14 the G-function of a Frobenius manifold is used to study the variance of the spectral numbers of a singularity. The construction of Frobenius manifolds in singularity theory is carried out in chapter 11. It requires the majority of the results which are presented in chapters 7 to 10. In chapters 7 and 8 a lot of material on meromorphic connections is given, most of which is known, but not presented in this form in the literature. It will be used for the discussion of the meromorphic connections in chapter 9 and 10. In chapter 10 most of the known results on the GauB-Manin connection for (unfoldings of) singularities are put together in a concise survey. In chapter 9 Frobenius manifolds are defined and certain meromorphic connections, which arise from them, are studied. This extends some work of K. Saito, Dubrovin, and Manin. Sabbah generalized most of K. Saito and M. Saito's construction to the case of tame functions with isolated singularities on affine manifolds [Sab3][NS] [Sab2][Sab4]. But the details are quite different; there one uses oscillating integrals, and the results are not as complete as in the local case. The case of tame functions is important for mirror symmetry. A special case had been studied by Barannikov [Ba3]. All of this is discussed in sections 11.3 and 11.4. 6.1 Construction of Frobenius manifolds for singularities A Frobenius manifold here is a complex manifold M with a multiplication о and a metric g on the holomorphic tangent bundle TM and with two global vector fields, the unit field e and the Euler field E. The multiplication is commutative and associative on each tangent space, the metric is flat, and all the data satisfy a number of natural compatibility conditions (see Definition 9.1). Let / : (Cn+', 0) -» (C, 0) be a holomorphic function germ with an isolated singularity at 0. The dimension of its Jacobi algebra 0c"+',o/(|f-. • • • > Щ-) ='¦ O/Jf is the Milnor number /л. A semiuniversal unfolding is a function germ 6.1 Construction of Frobenius manifolds for singularities 101 f ; (C"+1 x O\ 0) -»¦ (C, 0) with F|(C"+1 x {0}, 0) = / and such that the derivatives |f |(C"+1 x {0}, 0), г = 1,..., ц, represent a basis of the Jacobi algebra; hereV t) = (*„, ...,*„, f,,.... rM) e Cn+1 x C. One can choose a representative F : X -» Д with Д = B\ С С, М = В? С С and X = F-'(A)n(B"+1 х М) С С"+1хС for suitable small s, S, 9 > 0. One should see it as a family of functions F, : X П (Ben+1 x {t}) -» Л, para- parameterized by t 6 M, with Fo = /. The manifold M is the candidate for a Frobenius manifold. Its tangent bundle carries a canonical multiplication: The critical space С С A" of the unfolding is defined by the ideal 7F = (Ц,.... f?). Its pro- projection ргс,м ¦¦ С -> M is finite and flat of degree /г. There is a canonical isomorphism Э — 3F F.1) of free Ом-modules of rank д, the Kodaira-Spencer map. Here TM is the holomorphic tangent sheaf of M. The natural multiplication on the right hand side induces a multiplication о on Тм- This was first observed by K. Saito [SK6][SK9]. With e = a-'(llc) as the unit field and E := a~\F\c) as the Euler field, M carries a canonical structure of an F-manifold (M, о, е, Е) with Euler field (Definition 2.8 and Theorem 5.3). The Kodaira-Spencer map gives for each tangent space T,M an isomorphism ]M,o,E\,)=( (? \xeSin (T,M Jacobi algebra of (F(, x), mult., [Ft]). F.2) In order to find a metric one may look for a similar isomorphism as in F.1) from the tangent sheaf to a sheaf with a nondegenerate symmetric bilinear form. Such a sheaf exists, the sheaf пх+/А,М F-3) of relative differential forms with respect to the map <p in F.5) is a free Ом-module of rank \x and is equipped with the Grothendieck residue pairing Jf : Qf x Qf -*¦ Ом (see section 10.4) It is also a free (prc,jw)*C>c-module of rank 1; generators are represented by suitable volume forms unit(x, f)djco... dxn. The choice of a generator induces isomorphisms (ргс,м)*Ос F.4)
102 Introduction to part 2 and the pairing Jf yields a metric on T M. But one needs special volume forms, the primitive forms of K. Saito, in order to obtain flat metrics and Frobenius manifold structures on M. Their construction uses the GauB-Manin connection. In the following we intend to give an idea of this construction. The details can be found in section 11.1. The map cp:X-+AxM, (x, t) ь» (F(x, t), t) F.5) is a C°°-fibration of Milnor fibres outside of the discriminant V := <p(C) с A x M. The cohomology bundle H"= (J Hn(q>-\z,t),C) F-6) (z.t)eAxM-i) has rank /г and is flat. There is a distinguished extension 7i@) of its sheaf of holomorphic sections over the discriminant. The sheaf Ti^ can be defined in two ways. In terms close to the relative de Rham cohomology it is It is a coherent and even free Од хМ -module of rank \x [Gre]. It has a logarith- logarithmic pole along V, that means, the derivatives of sections in 7i@) by logarithmic vector fields are still in 7i@). The residue endomorphism along Vreg has eigen- eigenvalues (^, 0,..., 0) and is (for n > 2) semisimple. The second description of 7i@) is this: it is the maximal coherent extension of the sheaf of holomorphic sections of H" over V with these properties along V (see Lemma 10.2 and section 10.3). One can extend the cohomology bundle H" uniquely to a flat bundle over С x M - V and the sheaf 7i@) to С х М. Now the key point is to look for extensions Wa) of 7i@) over P'xM which are free OPi x ^-modules and which have a logarithmic pole along {oo} x M. Figure 6.1 may help to visualize the situation. log. pole {oo} x M A x {0} log. pole_ Figure 6.1 6.1 Construction of Frobenius manifolds for singularities 103 Extensions with a logarithmic pole are not difficult to obtain. Denote by H°° the д-dimensional space of the global flat manyvalued sections in (C - A) x M of the extended cohomology bundle (there is only the monodromy around {oo} x M). This space Я°° is equipped with a monodromy operator. There is a one-to-one correspondence between locally free extensions W0) with a logarithmic pole along {oo} x M and monodromy invariant (increasing exhaustive) filtrations U, on H°° (see sections 7.3 and 8.2). But for a free Opi xM-module H*® one needs special filtrations U.. First one observes that it is sufficient to show that the restrictions of the sections to P1 x {0} yield a free OPi-module. A classical theorem (cf. for example [Sab4,1 5.b], [Mal4, §4]) on families of vector bundles over P1 then asserts that the sheaf over P' x M is a free module (for M sufficiently small). The germs in (A x {0}, 0) of the sections in 7i@) form the Brieskorn lattice Hq [Bri2], a free C{z}-module of rank /z. In general it does not have a logarith- logarithmic pole at 0, but its sections have moderate growth, so it is regular singular. Varchenko showed that the principal parts of its sections give rise to a Hodge filtration on Я°° (see section 10.6). His construction was modified to obtain Steenbrink's Hodge filtration F' on H°°. Here Я°° is canonically identified with the space of the global flat multivalued sections on the cohomology bundle over A* x {0}. Now M. Saito found that an opposite filtration U. to this Hodge filtration (see Definition 10.19) is what one needs. In fact, he did not look for extensions to oo, but he constructed from U, a basis for Щ with properties such that he could apply Malgrange's results. The fact that an opposite filtration U. gives rise to an Opt -free extension to P1 of the Brieskorn lattice Щ with a logarithmic pole at oo is a solution of a Riemann-Hilbert-Birkhoff problem and is discussed in section 7.4. The existence of opposite filtrations U. to F' follows from properties of mixed Hodge structures. In general U. is not unique. Now fix a choice of U. and the corresponding extension 7i@). Denote by n : P1 x M —у M the projection. The sheaf я,7^@) of fibrewise global sections is a free Ом -module of rank д. It contains the /x-dimensional subspace of sections whose restrictions to {oo} x M are flat with respect to the residual connection along {oo} x M (see section 8.2). The residue endomorphism along {oo} x M acts on this space. It turns out that it acts semisimple with eigenvalues — скь ..., —aM (the rational numbers a\,..., aM are the spectral numbers of Щ, see sections 7.2 and 10.6). The smallest spectral number or i has multiplicity 1 (Theorem 10.33). Let v\ be a global section on 7i@) which is flat along {oo} x M with respect to the residual connection there and which is an eigenvector with eigenvalue
104 Introduction to part 2 —a\ of the residue endomorphism. It is uniquely determined up to a scalar. It turns out that it is a primitive form in the sense of K. Saito. It yields a period map and an isomorphism л*Н@) = TM. The residual connection along {00} x M induces a flat connection on TM. In order to see that the isomorphism F.4) for Vi as volume form gives a flat metric on TM one needs two things: K. Saito's higher residue pairings (see section 10.4) and the fact that the Hodge filtration F* on H°° is part of a polarized mixed Hodge structure. There is a polarizing form S on H°° which also has to be respected by the opposite filtration U.. It is related to the higher residue pairings (see section 10.6). Altogether one obtains the following (see Theorem 11.1). Theorem 6.1 A monodromy invariant opposite filtration U. to the Hodge fil- filtration F' on #°° induces a flat metric g up to a scalar on M such that (А/, о, е, Е, g) is a Frobenius manifold. All the material in Sabbah's book [Sab4] and the second structure connections in Manin's book [Man2, II§2] have been very helpful for carving the above version of the construction of Frobenius manifolds in singularity theory. The second structure connections in [Man2] are a family Vw, s e C, of meromorphic connections over P1 x M for a semisimple Frobenius manifold. The definition generalizes to arbitrary Frobenius manifolds (chapter 9). In the case of singularities the connection V(~l' turns out to be isomorphic to the (extended) GauB-Manin connection with the sheaf 7i@\ The connection V@) was also defined by Dubrovin [Du3]. In the singularity case, K. Saito had already defined the germs at 0 of all the connections V(i> in a different way [SK9, §5]. 6.2 Moduli spaces and other applications Singularities which are contained in a д-constant family have isomorphic Milnor lattices and for и ф2 even the same topological type. They are called /x-homotopic. Singularities which differ only by a local coordinate change are called right equivalent and should be considered as isomorphic. One may ask about the moduli space of right equivalence classes of singularities in one ц- homotopy class. Let us fix a /x-homotopy class ? с m2 с Ос+'.о of singularities and an integer к > /x + 1. The right equivalence class of a singularity / € ? is deter- determined by its &-jet jkf e m2/mi+1. The set jk? с m2/m*+1 is a quasiaffine variety (possibly reducible as a variety, but connected as a topological space). The algebraic group jk7l of ?-jets of coordinate changes acts on it. The quotient parametrizes the right equivalence classes of singularities in the 6.2 Moduli spaces and other applications 105 /x-homotopy class ?. One knows that the orbits all have the same dimension, but a priori not much more about the group action. We can prove the following (see Theorem 13.15). Theorem 6.2 The quotient ]к?/]кТ1 is an analytic geometric quotient. A priori this is a global statement. But with the construction of unfoldings and with some results of Gabrielov and Teissier one can translate it into statements on semiuniversal unfoldings of singularities in ?. Then one can use the rich structure of their base spaces as Frobenius manifolds. Theorem 6.2 includes the claim that the quotient topology on jk?/jk1l is Hausdorff. I now want to sketch the proof of that part in one page. Consider a singularity / e ? and a semiuniversal unfolding F with base space M as in section 6.1. The /x-constant stratum in M is S^ = {t € M | Sing F, = {x}zndF,(x) = 0}. A result of Scherk [Sche2] says that for any t e M the datum in F.2) determines the right equivalence classes of the germs (F,, x) for x e Sing F,. So the base space as F-manifold (M, о, е, Е) with Euler field knows the right equivalence classes of all the singularities above it. There is a related result of Arnold, Hormander, and others on Lagrange maps and generating families. It implies (Theorem 5.6) that the germ ((M, t), о, е, Е) of an F-manifold for t G 5M determines the germ of the unfolding at the singular point of F, up to right equivalence of unfoldings. Now consider for /, F, M, and 5Д(/) as above a sequence (r,-)ieN with U e SM(/)andr,- -> Ofori -*¦ со and suppose that there is a second singularity / with F, M, 5Д(/) and (?¦); defined analogously and with a sequence of coordinate changes <p, such that Fu — F?. о щ. One has to show that then / and / are also right equivalent. This will imply that the quotient topology is Hausdorff. Figure 6.2 illustrates the situation. There is no possibility of controlling the coordinate changes <p, and finding a limit coordinate change. But it turns out that they induce unique isomorphisms <PiM ¦¦ , t,), о, е, Е) -»• i), о, е, Е) F.8) of the germs of F-manifolds. With the construction of Frobenius manifolds we can control these and show that there is a subsequence^which gets stationary for large (' and extends to an isomorphism (M, 0) -> (M, 0). Then опекал apply Scherk's or Arnold's and Hormander's result and see that / and f are right equivalent. Essential for controlling the sequence <pi<M is the strong link between flat structures on M and filtrations on H°° by Theorem 6.1. The coordinate changes
106 Introduction to part 2 M Э M D Figure 6.2 <Pi also induce isomorphisms <pUtoh : (Я00, Я2°°, Л, S, F'fo)) -+ (Я --, ^z- here Я00 is the space from section 6.1 with canonical lattice Я^°, monodromy operator h, polarizing form S and Hodge filtration F*(r,) from the singularity Ft,- The data (Я00, Я|°, Й, 5, F»(/)) encode polarized mixed Hodge structures. There exists a classifying space DPMHS for them and a period map Sn ->¦ DPMHS, t м- F\t). F.10) Я~, /Г, F.9) The discrete group Gz = Aut(ff°°, Я^°, h, S) acts properly discontinuously on DPMHS (because the mixed Hodge structures are polarized). This implies that the isomorphisms <Pi,coh are all contained in a finite set. Via the construction of flat structures on M the same then holds for the isomorphisms (pi,m (for more details see section 13.3). One can be more precise about the local structure of the moduli space jke/jkil- The group of automorphisms Aut((M, 0), о, е, Е) is in fact finite and acts on the /x-constant stratum 5M С М. Similar arguments as above yield the following for a germ of the moduli space (see Theorem 13.15). Theorem 6.3 Uk?/JkK, [jkf]) S (?„, 0)/Aut((M, 0), о, е, Е). We can equip the /x-constant stratum with a canonical complex structure. An opposite filtration U. as in Theorem 6.1 induces a flat metric up to a scalar on 6.2 Moduli spaces and other applications 107 U and a unique flat structure. By construction there exists a basis of flat vector fields Si, .--.iSm with The coefficients su e in F.П) F.12) are close to a part of Dubrovin's deformed flat coordinates (see section 11.2) and they determine the //.-constant stratum (see section 12.1). Theorem 6.4 The ^-constantstratum (Stf, 0) is the zero set of the ideal (e,j | aj — 1 — a.; < 0). This ideal is independent of the choice of Si,..., SM and even of the choice of the opposite filtration U.. I expect that this ideal provides in general a nonreduced complex structure. But it is very difficult to compute. I do not have examples. I also expect that the induced complex structure on the moduli space jk^/jk^l is a good candidate for a coarse moduli space with respect to some functor of /x-constant deformations over arbitrary bases. In [Hel][He2][He3] Torelli type questions for hypersurface singularities were studied. There is a datum which is even finer than the polarized mixed Hodge structure (Я°°, H$°, h, S, F*)of a singularity, the datum (H°°, H™, h, S, Щ), the Brieskorn lattice together with topological information. A classify- classifying space Dbl for such data was constructed in [He4]. It is a fibre bundle over the classifying space DPMhs- The group Gz also acts properly discontinuously on Dbl- There is a local period map D BL, F.13) for the /x-constant stratum of a singularity and a global period map DBL/GZ F.14) for a /x-homotopy class of singularities. A global Torelli type conjecture [He2] asks whether F.14) is always injective. In all known examples it is true, but a general answer is still unknown. Now Theorem 6.2 shows at least that the moduli space on the left is an analytic variety. The map F.14) is now a morphism between varieties.
108 Introduction to part 2 It is easy to see that F.13) is injective (for small 5M) if 5M is smooth. M. Saito [SM4] used this fact to show that for general 5M it is finite-to-one. (He did not have the classifying space Dbl, but considered a period map to a bigger space, a subset of a certain flag manifold.) With the flat coordinates of a Frobenius manifold structure on the base space M we can show a stronger infinitesimal Torelli type result (see Theorem 12.8). Theorem 6.5 The period map 5Д -*¦ Dbl is an embedding for small S^. All these applications made use of the fiat structure on M which is induced by an opposite filtration, but not of the metric itself and not much use of its other properties, the multiplication invariance and the potentiality (see Definition 9.1). These give rise to an extremely rich hidden structure on Frobenius manifolds which has been uncovered by Dubrovin and Zhang [DuZl ] [DuZ2] and Givental [Gi8]. Exploiting this for singularities is a big task for the future. We can present one surprising application. There is a function G{t) associated to each semisimple Frobenius manifold ([DuZl][DuZ2][Gi7], cf. section 14.2). Following a suggestion of Givental we show (Theorem 14.6) that it extends in the singularity case to a holomorphic function on the whole base space. It has the stunning property 48 =: y. F.15) The spectral numbers a,- satisfy the symmetry a,- + a^+i_,- = n — 1. So one can consider <~ as their expectation value. Then - JX^a,- - ^J is their variance. The G-function gives a grip at this variance. In the case of a quasiho- mogeneous singularity /, one has / e J/ and E\o = 0 because of F.2). This shows the following (Theorem 14.9). Theorem 6.6 Iff is a quasihomogeneous singularity then у = 0. The variance of the spectral numbers is 2 ) VI It is part of the motivation for the following conjecture. F.16) Conjecture 6.7 For any isolated hypersurface singularity, the variance of the spectral numbers satisfies 12 F.17) Chapter 7 Connections over the punctured plane The only initial datum in section 7.1 is a monodromy operator. For the corre- corresponding flat vector bundle over the punctured plane C* notions such as the elementary sections and the V-filtration are introduced. In section 7.2 Oc-free extensions over 0 with regular singularity at 0 of the sheaf of holomorphic sections of the vector bundle are discussed. Comparison with the V-filtration leads to the spectral numbers and certain filtrations. Sections 7.1 and 7.2 are elementary and classical. The subject of section 7.3 is those extensions over 0 which not only have a regular singularity at 0, but also a logarithmic pole. There is a correspondence between such extensions and certain filtrations, which is not so well known. It has a generalization in section 8.2. It is used in section 7.4 for the solution of a Riemann-Hilbert-Birkhoff problem. This is based on ideas of M. Saito. It is central to the construction of Frobenius manifolds in section 11.1. In section 7.5 a formula is given for the sum of the spectral numbers in a global situation, when one has on a compact Riemann surface a locally free sheaf and a flat connection with several singularities as above. It is useful in the case of P1 for the Riemann-Hilbert-Birkhoff problem. 7.1 Flat vector bundles on the punctured plane Let us fix a holomorphic vector bundle H -*¦ C* of rank д > 1 with a flat connection V on the punctured plane С* = С — {0}. We want to discuss special sections in H and extensions of the sheaf H of its holomorphic sections over 0 6 C. Of course this is classical and has been done in many ways (e.g. [Del], [AGV2], [SM3], [Hel], [Ku], [Sab4]). But we have to establish comfortable notations in order to discuss later the information which is contained in certain 109
по Connections over the punctured plane extensions of H over 0, for example in the Brieskorn lattice (section 10.6). We more or less follow [SM3][Hel]. A positively oriented loop around 0 induces a monodromy h on each fibre Hz, z € C*, of the bundle. The monodromy determines the bundle uniquely up to isomorphism. Let h = hs- hu = hu- hs be the decomposition into semisimple and unipotent parts, N :— log hu the nilpotent part, and := кег((/ь - A.): Hz -+ Hz), G.1) the decomposition into generalized eigenspaces. We will use the universal covering e:C->C*, f н> e2jrlf. G.2) Global flat sections A of the bundle e*H -у С induce via the projection pr : e*H -*¦ H maps pr о А : С -*¦ H, which are called global flat multivalued sections. The space of these global flat multivalued sections is denoted by H°°. It is canonically isomorphic to each fibre (e*H)(, f e C. The monodromy h acts on it with eigenspace decomposition H°° = 0Л Н?° and with h A(f) = A(r + 1) * G.3) for any Л б Н°°. Now we can define some special global sections in H. Fix A e H™ and a € С with enia - X. The map С ^ H, fn e(af) exp(-f /V)A(f), is invariant with respect to the shift f i-> f + 1 and therefore induces a holo- morphic section es(A,a) : C* -»¦ Я, fore(f) = г G.4) of the bundle #. It is called an elementary section [AGV2] and is usually denoted informally as / N \ zaexp -logz- —• A, V 2л-(/ n G.5) a is called its order. It is nowhere vanishing if А ф 0 because the twist with e(af) exp(—f TV) is invertible. The symbol Ca denotes the space of all elementary sections es(A, a), A e H™, with a fixed order a. The map фа : Ял°° -> С", es(A, a) G.6) 7. / Flat vector bundles on the punctured plane 111 is an isomorphism of vector spaces. By definition one has z -es(A,a) = es(A, a + 1), z о т/га = фа+\, ' G.7) (-N \ Va,es(A, a) = a ¦ es(A, a - 1) + es ( ^—7A, a - 1 I, !,«), / N \ N (zV8 - a)e5(A, a) = es --—7A, a I = --—r \ Zjvi / Zni G.8) z • ca ->¦ Ca+1 bijective, V3z: С -> С" bijective iffa^O, zV3 - a = : Ca -»¦ С nilpotent, 2 2л i so C" is a generalized eigenspace of zV8;. We simply call these spaces С To obtain a filtration for these eigenspaces we fix a total order -< on the set {a | e~2jtia eigenvalue of h] U Z, which satisfies a -< a + 1, a4|S iff a + 1 ч /3 + 1, Va, ^Э/neZ a < P+m. G.9) Later h will be quasiunipotent and, if not said otherwise, the order will be the natural order < on Q. But different orders can also be interesting (cf. Remarks 11.7). To simplify notations, we will write the usual symbol < for the order <. It should be clear when the usual order < on E is meant and when the order -< on [a | е~1л1а eigenvalue of h} U Z is meant. From now on we will concentrate on germs at 0 of sections in H, that means, on the stalk (i*H)o at 0 of the sheaf ?«H, where / : C* «-> С is the inclusion. The space of function germs (i*Oc«)o and the operator Vaj act on this stalk. The eigenspaces Ca and the elementary sections are identified with their images in o. Then C" is characterized as Ca = ker((zVa2 - a) G.10) for some m 3> 0. A basis of elementary sections in ©_i<a<o Ca induces а basis of each fibre Hz, z e C*. Therefore = ?ft (i.Oc.)oC" G.11) is an (i,0c')o-vector space of dimension ti = dime Яг (г € С*).
112 Connections over the punctured plane The space of all germs at 0 of sections of moderate growth (in the sense of Deligne [Del, II §§1-4], comparing them with the flat multivalued sections) is G.12) The space V> °° is a C{z}[z ']-vector space of dimension ix. It is a regular holonomic ?>c,o-module of meromorphic type [SM3, § 1]. Obviously it decom- decomposes into T>c,o- and C{z} [z~' ]-submodules which correspond to a Jordan block decomposition of H°°. Within this space one has the Kashiwara-Malgrange V-filtration, an ex- exhaustive decreasing filtration, indexed by {{a \ e"ia eigenvalue of h), <) and defined by C{z}Cfi. One supplements this with V>a := 0 G.13) G.14) Then V* and V>ff are free C{z}-modules of rank д, satisfying z : у" -» Va+1 bijective, G.15) V3; : V Va. : V a » V" >0 bijective if or > 0, bijective, Gr% := Va/ V>a = Ca canonically isomorphic. G.16) Any section cd e V>-°° is a sum (often infinite) of unique elementary sections, <w = ]T ¦$(«,«), s((o,a)eC, G.17) a whose orders are bounded from below by some number or(«) := max(a | a> 6 V") = min(a | s(co, а) ф 0), G.18) which is called the order of a>. The elementary section s(w, ct(co)) is the principal part of со. All the sections s(co, a) are called the elementary parts G.19) Finally, we will need the ring 7.2 Lattices 113 of microdifferential operators with constant coefficients [Phi, part 2]. Just as Clz} itIS a discrete valuation ring and a principal ideal domain. In view of the action 3r'z* = шг*+' of 9^' on C{z}> the ring C{{9^'11 is designed to act on C[z) such that C{z} is a free С{{Э-'}}-тос1и1е of rank 1 with generator 1. It is well known ([Mal2, 4.1]) (and not hard to prove elementarily) that this generalizes as follows. Lemma7.1 ThemapVd. : C{z}Co+1 ->• C{z}Cafora i Z<0 is bijective, the inverse extends to an action ofC{{d^}} such that C{z}Ca is a free C{{d~1}}- module of rank dim C™. Especially, V*~l and all V, V>a for a > -1 are free C[[d;l}}-modules 3 7.2 Lattices We stay in the situation of section 7.1. Up to now the only initial datum was the monodromy h. It determined the flat bundle H —>¦ C* up to isomorphy. Everything in section 7.1 was developed from this. But usually one has another ingredient, a C{z}-module in V>-o°, which contains additional information (e.g. the Brieskorn lattice, cf. section 10.6). One wants to understand this information by comparing with the V-filtration and the elementary sections. We first discuss C{z]-modules of У>~°0, most of the discussion also applies to Qta-'H-submodules of V>. A finitely generated C{z]-submodule of У >~°° is free, as C{z] is a principal ideal domain. The name C{z] -lattice will be reserved for free C{z}-modules of the maximal rank /x. A C{z]-lattice ?o can be extended uniquely to an Ос-free subsheaf of rank ?i in i^H. The correspondence between C{z]-lattices Co С V>~°° and Oc-free subsheaves of rank д of i^H whose sections have moderate growth is one-to-one and justifies focusing on the stalks at 0. A C{z}-lattice ?0 satisfies C{z][z~']?o = V>-oo,justasdotheC{z]-lattices V". Therefore there exist a' and a" with Va' D C0D Va". The principal parts of the sections of Co are placed together in the subspaces = (V" П Co + V>e)/V>- С Or" = C. G.20) One can visualize them in figure 7.1. Obviously zGry"' Cq = Gryz?o С Gry?o. The dimensions of the quotients are dim 1 Co) = dim Gr^?0 - dim Gi^' ?0 =: d(a) G.21)
114 Connections over the punctured plane Са+Ъ -1 V' a + 1 ¦>o a +2 a + 3 Figure 7.1 and satisfy G.22) keZ They give rise to the spectral numbers (<x\,..., a^) = Sp(Co) of the lattice Co [AGV2], defined by tt(/ | a,- = a) = d(a), G.23) and ordered by a; < • ¦ • < aM. Via the isomorphisms i/ra : H?° —> C" (cf. G.6)), the subspaces Gry?o induce an increasing exhaustive uj-invariant filtration F. on H°°, for Я = , -1 < a < О, FPH°° := G.24) G.25) Remarks 7.2 (a) In the V-filtration, in ai < • • • < a^, and in —1 < a < 0 the order -< from G.9) is used. (b) In the case of the Brieskorn lattice of an isolated hypersurface singular- singularity /(*o. ¦ ¦ ¦, xn), the filtration F* = Fn_. is essentially Varchenko's Hodge filtration ([Val], cf. also section 10.6). It reflects the information contained in the principal parts of the sections in the Brieskorn lattice and is already highly trancendental. The question of how to treat the higher elementary parts of the sections and the whole information in the Brieskorn lattice leads to M. Saito's 7.2 Lattices 115 work [SM3, §3], which will be taken up in section 7.4, and to my Torelli type results ([Hel]-[He4] and section 12.2). Lemma 7.3 A C{z}-lattice Co С V>~°° with spectral numbers a\ < • • • < aM satisfies Vе" dC0D V>a"-'. G.26) Elements u>\,..., a>M e Co whose principal parts represent a basis of the space 1 Co form a C{z}-basis of Co- Proof. We may suppose a(&>;) = a,-. The elements &>ь ..., &>д areC{z}-linearly independent because in any linear combination with nonvanishing coefficients at least one of the principal parts is not cancelled by anything. They generate a C{z}-lattice?Q С Со- Because of G^C'q = Gr^?o for all a, one can enlarge the order of elements of Co arbitrarily by adding elements of C'o. So Cq с C'o+V^ for some large ft and then Co = C'o. In the same way one obtains У>а^ с С'о + Vp = C'o = Cq. The inclusion V Э Co is obvious. ? AfreeCKa-'H-modulero С V>- 'of maximal rank /miscalled a C{{d-1}}- lattice. Spectral numbers ai,..., aM are defined as in G.21) and G.23). One obtains an increasing exhaustive /ij-invariant filtration F"lg on H°° by Ff*H? = if'1 v?Grv+p?o for X = e™, -1 < a < 0, G.27) palg яоо _ д^ palg яоо (? 2g) Lemma 7.3 holds analogously for C{{dz '^-lattices. Often a subspace Co С V>-' is given which is a C{z}- and a C{{Зг"' }}-lattice. Then the definitions G.21) and G.23) give the same set of spectral numbers when one replaces z in G.21) by 3. But the filtrationsF. and Ffg may differ. Lemma 7.4 (a) A C{z}-lattice Co С V>'1 is a C{{d-l}}-lattice if and only if V-!?o С Co. A C{[d-1}} -lattice ?0 С V2" is a C{z}-lattice if and only if zC0 С Co. (b) The filiations F. and F?g of a C{z}- and C{{d-l])-lattice Co С V>~1 satisfy F",g = F_i =0 and are related for p > 0 by the formula F°plgH? = П (-?j + к + a") FpH where X = е'2*'", -1 < a < 0. 1.29)
116 Connections over the punctured plane Г 7.3 Saturated lattices 117 Proof: (a) The inclusion ?Q Э V' for some a', Lemma 7.1, and Va l?0 С ?0 imply that ?0 is a C{{3~'}}-lattice, analogously when z?o С ?q for C{z}. (b) This follows from the definitions of F. and F^lg and from the formula 2Va2 - a = -? on Ca (cf. G.8)). D Remarks 7.5 (a) The filtration Fp g has the index 'alg' because in the case of a hypersurface singularity f(xo,.... xn) the operator V8. has a more algebraic flavour than the operator z (cf. sections 10.3 and 10.6). (b) The Brieskorn lattice is a C{C~' }}-lattice, and F°'j. is the Hodge filtra- filtration of Steenbrink [Stn] (and Scherk [SchSt], M. Saito [SMI], Pham [Ph3]). Because of G.29) the filiations F°'l. and F,,_. coincide on the quotients of the weight filtration (which is defined via N, cf. [Schm][AGV2] and section 10.5) and are both Hodge filtrations of mixed Hodge structures. But F^J, behaves better with respect to a polarizing form and is part of a polarized mixed Hodge structure ([He4], cf. sections 10.5 and 10.6). (c) This polarizing form is, after the monodromy, a second topological ingre- ingredient and can be married to the structure {V~l, C, z, 9~') giving aC{{3~'}}- sesquilinearformon V, which in fact coincides with the restriction to y>-1 of K. Saito's higher residue pairings and which fits together with the Brieskorn lattice [He4]. This will be discussed in section 10.6. 7.3 Saturated lattices We stay in the situation of section 7.1. A saturated lattice ?q is a free C{z}- module ?q С (i*H)o of rank /z with z V3.,?o С ?q. Then the germs of sections in ?o have moderate growth, that means, ?o С V>~°°. This follows from the classical theorem of Sauvage, that the solutions of a system of linear differential equations with simple pole have moderate growth (cf. for example [Del, II. 1] or [Sab4, II.2]). We will see that there are correspondences saturated lattices ¦o- filtrations on H -*¦ C* by flat subbundles -o- monodromy invariant filtrations on #°° G.30) and that the saturated lattices and these correspondences are invariant under change of the coordinate z. They are even independent of the choice of the order < in G.9). A result related to these correspondences has been given by Sabbah[Sab4,IIIl.l]. The structure of a saturated lattice is summarized in the following lemma. Lemma 7.6 (a) Let ?Q С V>~°° be a saturated lattice with spectral numbers a,,.. •, aM (cf. G.23)j and filtration F. on H°° (cf. G.24) and G.25)). Then G.31) so ?o contains all the elementary parts s(co, a) of a section со s ?q. The spaces ?0ПСо С С and Fp С H°° are N-invariant and thus monodromy invariant. (b) Any increasing exhaustive monodromy invariant filtration F. on Я°° induces a saturated lattice in the following way. The filtration F. on H°° induces a filtration of H -> C* by flat subbundles FpH —> C*; the sheaf of holomorphic sections inthe subbundle FpH isdenoted by FpH. Then " G.32) is a saturated lattice and F. on H°° is the filtration of?0 defined in G.24) and G.25). Proof, (a) Lemma 7.3 yields the inclusions Vе" D ?o Э V>a»-1. The oper- operator zV3; acts on the space yei/y>ov~' = ®a,<ff<a(l_iCa, its generalized eigenspaces are the spaces C". The subspace ?qIV>"ii~1 is invariant un- under zV3j. This implies G.31) and zV8z(?0 П С") С^П С". The formula z V8j - a = -?, on Ca (cf. G.8)) gives the ЛГ-invariance of ?0 П С and of Fp С Я°°. (b) Exhaustive means that there exist integers a < b with 0 = Fo С fj = H°°. Then V>a D ?0 Э V>b~l. The stalk (i*FpHH is an («„OcOo-module and invariant under Vs.. The C{z}-module <C{z\Ca+p is invariant under zV8t. Therefore ?o is a saturated lattice. The equalities Grv+P?o = ?оП C+p = (i,FpH)o П Ca+P G.33) for -1 < a < 0 show that G.24) and G.25) give the filtration F. on H°°. О If <p : (C, 0) -> (C, 0) is an isomorphism of germs then <p(z) itself can be considered as a new coordinate on (C, 0). One can lift <p to an automorphism Ф of the flat bundle H -»• C* (in a neighbourhood of 0) such that Ф : Hz -» Нф) is an isomorphism obtained from the flat structure by some path from z to (p(z). The isomorphism Ф is not unique, only unique up to a power of the monodromy. The isomorphism Ф maps (i*H)o to itself and C{z}-lattices to С[z]-lattices.
118 Connections over the punctured plane But Ф(С") does not usually coincide with Ca: for the definition of elementary sections the choice of a coordinate z was crucial. Also many C{z)-lattices are not invariant under Ф. Lemma 7.7 Let <p and Ф be as above. Then the sets C{z}Ca and all saturated lattices are invariant under Ф. Proof: One can use G.5) to see that an elementary section is mapped to a section (not necessarily elementary) with the same order. Thus Ф(С°) С C{z}Ca, and Ф-' gives Ф(С{г}Са) = C{z}Ca. Now formula G.32) shows that any saturated lattice is invariant under Ф, because the flat subbundles are invariant by definition of Ф. D The subspaces V" of the Kashiwara-Malgrange V-filtration are saturated lattices. Their independence of coordinates is of course known. We could have considered the spaces C{z]Ca as the subspaces of the V-filtration for the flat eigenspace subbundle ket(hs - е~1жш) -> С* with eigenvalue e"l7tia. Now, for a coordinate free reformulation of the previous discussion, let M be a 1-dimensional manifold, q e M a point, i : M — {q} e-> M the inclusion, and H -» M - {q} a flat vector bundle with sheaf Ti of holomorphic sections. The OM,q-modules Cq С (itH)q of rank д (= rank of H -+ M - {q}) are called Ом,ч-lattices. They are the germs in q of CM-locally free extensions С of H over q. If z is a coordinate around q then the vector field zdz generates the logarithmic vector fields in a neighbourhood of q (cf. section 8.2). We can rewrite the above definition of a saturated lattice coordinate freely. Definition 7.8 An (9M,9-lattice Cq С (i*H)q is saturated if and only if it is invariant under the logarithmic vector fields along [q] С М (cf. section 8.2). An CV-locally free extension CofH over q has a logarithmic pole at q if and only if the lattice Lq is saturated. The Lemmata 7.6 and 7.7 give the following. Theorem 7.9 The saturated lattices correspond one-to-one to the increasing exhaustive filtrations by flat subbundles of the restriction of H to some disc around q. 7.3 Saturated lattices 119 A saturated lattice Cq С (i*H)o is equipped with a residue endomorphism Res, = zVs. Cq/zCq Cq/zCq G.34) here z € Ом,ч is a coordinate with z(q) = 0. The residue Res(/ is independent ofthe coordinate: if? is any other coordinate with ?(<?) = Ofhenz^ = u(z)zdz with и eOj,,,,u@)= 1. In fact, the vector space Cq/zCq can be identified with the fibre at q of a vector bundle on M which extends H and has Cq as the space of germs of holomorphic sections. Then the monodromy automorphism h of the bundle H (in a neighbourhood of q) extends to this fibre and coincides there with e-2lr'Res« [Del, II 1.17]. The endomorphism e-2l"Resi has the same eigenvalues as h, but may have a simpler Jordan block structure. The following more precise (and well known) statements will be useful later. We again fix a coordinate z, identifying (M, q) and (C, 0) and having the spaces С of elementary sections at our disposal. Theorem 7.10 LetC0 С (itH)obe a saturated lattice and F.the corresponding (mondromy invariant) filtration on H°°. (a) The coordinate z induces an isomorphism ААА^фОгРН00 G.35) p which identifies the actions 0fe-27riReso On the left hand side and of h on the right hand side. (b) The eigenvalues of the residue endomorphism Reso are the spectral numbers cc\,..., aM of Co. (c) The endomorphism Res0 is semisimple if and only if N(FP) С Fp-\ for all p. (d) The endomorphism e-27r'Res» has the same Jordan normal form as h if no two spectral numbers differ by a nonzero integer (nonresonance condition). Proof, (a) The coordinate z provides the spaces С of elementary sections. Because of G.31) and G.24) this yields isomorphisms — 1<а<0 р G.36)
120 Connections over the punctured plane On each subspace Ca the monodromy h acts as exp (-2л iz Va.) (cf. G.8)). This shows (a). (b) and (c) follow immediately from (a) and its proof. (d) Under the nonresonance condition there exists for each eigenvalue X of h an index p(X) with Gr?wHf = Щ0. Now (d) also follows from (a). Q Example 7.11 Suppose that the residue endomorphism Res0 of a saturated lattice ?0 С (i»W)o is semisimple with eigenvalues (s, 0,..., 0) for some s €C. If s e С - (Z - {0}) then the monodromy is semisimple with eigenvalues («Г2*", 1,.... 1) and the lattice is ?0 = C{z}C° ф C{z]Cs. If 5 e Z - {0} there are two possibilities: (a) Either the monodromy is the identity; then the saturated lattice can be any one in a family parameterized by P C° for 5 < 0 and by VHom(C°, C) for s >0. (b) Or there is one 2x2 Jordan block; then the saturated lattice is unique; for s > 0itisC{z}ker(W : C° -» C°)+Vs,fors < OitisC{z}zs-N(C0)+V°. 7.4 Riemann-Hilbert-Birkhoff problem The most recent references in book form on the Riemann-Hilbert problem and the Birkhoff problem are [AB] and [Sab4]. In chapter IV in [Sab4] Sabbah discusses several versions of the problem. One unifying general version can be stated as follows. Hypotheses: E, С P1 and S2 С P1 are two disjoint finite sets, H -> P1 - (Si U ?2) is a flat vector bundle of rank /x > 1 and with sheaf H of holomorphic sections, and ? is a free 0pi-x;2-module of rank /x with an isomorphism?|P! - (Si U S2) = H. Problem: Does there exist an extension of ? to a free OV\ -module ? which has logarithmic poles along the points of S2? Because of Theorem 7.9 one can extend ? to a locally free CPi-module with logarithmic poles along E2 without any problem. The requirement that ? shall be a free СУ -module makes the problem difficult. Often the problem is formulated in terms of the (trivial) vector bundles which correspond to the sheaves ? and ?. Usually one makes additional assumptions on the poles along S i. At least one supposes that for q € S| the coefficients of the connection matrix with respect to a basis of Cq are meromorphic. A much stronger assumption would 7.4 Riemann-Hilbert-Birkhoff problem 121 be that the sections in Cq, q e Sb are of bounded growth, that means, the connection is regular singular there. The classical Riemann-Hilbert problem is the case Si = 0, the classical Birkhoff problem is the case S, = {0}, S2 = {oo}, with the assumption that the connection matrix with respect to a basis of ?0 has a pole of order < 2. Sabbah [Sab4] calls the case S i = {0}, S2 = {oo} without special assumptions on ?o the Riemann-Hilbert-Birkhoff problem. A particular case of it was treated implicitly by M. Saito [SM3, §3]. He gave (implicitly) a correspondence between certain solutions and certain filtrations on the flat bundle H —> C*. The purpose of this section is to resume and generalize this correspondence. In the case of hypersurface singularities such filtrations exist because of properties of mixed Hodge structures (cf. section 10.5). Let us stick to the situation and notations in section 7.1. Again H -» C* is a flat vector bundle of rank /x and with monodromy h — hs ¦ hu = hu ¦ hs, N = loghu, and П, Я°°, С, фа, Va, V>a, СЦд;1}} as in section 7.1. Again ?0 с V>~°° will be a C{z}-lattice (of rank д), and ? is the Ос-free, extension of H to 0 with germ ?0 at 0. Because of Theorem 7.9 the extensions of ? to locally free СУ -modules ? with a logarithmic pole at oo correspond one-to-one to the monodromy invariant finite increasing filtrations U. on H°°. Theorems 7.16 and 7.17 will describe distinguished extensions to free Opi -modules. The objects in section 7.1 for the point oo instead of 0 will all be equipped with a tilde: the coordinate ?= j, satisfying d? dz the spaces Ca = C~a, V>a С V С V>-°° a,,..., а„ for a C{z}-lattice ?ю С V>0. -zdz, G.37) ,7^H0, the spectral numbers Example 7.12 Let H -*¦ C* be the trivial bundle of rank 2 with basis e\, e2 of flat sections. The following is a 1-parameter family of free CPi -modules ?(r), r €<C, extending H to P1, ?(r) + ze2) • ze\. G.38) The lattice ?oo('-) = V is constant and saturated with spectral numbers Cfi, a2) = (—1, -1). The lattice ?0(r) С V° is not constant and saturated only for r = 0; even the spectral numbers jump, (ai,a2)@) = A,1) and (ai, a2) = @, 2) for г ф 0. A C{z}-basis of ?0(r) for г ф 0 whose principal
122 Connections over the punctured plane parts represent a basis of the space 0a Gr^.Co(r)/Gr^z.Co(r) (cf. Lemma 7.3) is given by C0{r) = C{z] ¦ {rex + zei) Ф C{z} ¦ (z2e2), гфО. G.39) It does not extend to a basis at r = 0. In the example J2i &i + Hi a> — 0- Theorem 7.20 will show that this holds in general. We will be interested in free Opi -modules with a stronger property, which includes a,- = —arM+i_,-. We resume two definitions from [SM3, §3]. Definition 7.13 (a) Let Co С V>~°° be a C{z}-lattice. A /x-dimensional sub- space W С Co such that the projection pr : W —>¦ Co/zCo is an isomorphism is the image of a unique section i> : Co/zCo —> Co with prov — id. The space W is called a good .Co/z-Co-section if the following two equivalent conditions hold: (i) The nitrations pr(V П W) and (V П?о + zCo)/zCo on C0/zC0 coincide, (ii) The space W has a basis whose principal parts represent a basis of the space (b)If?0 С V>-l isaCffa^Ml-lattice, good ^o/V^'A-sections are defined analogously. Remarks 7.14 ([SM3, §3]) A basis of the image Im v of any C-linear section v : Co/zCo -»¦ Co of a C{z}-lattice ?0 С V>-°° is a C{z}-basis of Co, by the lemma of Nakayama. Always pr(Va <1W)C (Va ПС0 + zC0)/zC0. In Example 7.12, the space W = С • (re, + ze2) + С • zex С Coif) is not a good ?o/z?o-section for г ф 0. The interplay between two filtrations on a finite dimensional space is dis- discussed in [De2]. If F. and U. are two increasing filtrations on a finite dimen- dimensional vector space then one has canonical isomorphisms 9Г Fp П П !/,_, + Fp_, П Uq) This helps to explain the equivalences in the following definition (cf. [De2], [SM3, §3]). Definition 7.15 Two increasing filtrations F. and U. on a finite dimensional vector space are called opposite to one another if the following three equivalent conditions hold: 7.4 Riemann—Hilbert—Вirkhoffproblem 123 (i) 0 then p + q = 0. (ii) The vector space splits into a direct sum 0p Fp П f/_p. (iii) One has decompositions Fp = ®q<p 1 for all p. ¦'qip Theorem 7.16 Let С be an Ос-free extension over 0 of the sheaf"H of sections of the flat vector bundle H -> C* with Co С V>~°°. The filtration which Co induces on H°° by G.25) is denoted by F.. There is a one-to-one correspondence between the two sets of data: (i) Extensions of С to Op-free modules С which have a logarithmic pole at oo and satisfy: the ix-dimensional space ?(P') of global sections is a good Col zCo-section, (ii) Monodromy invariant increasing exhaustive filtrations U. on Я°° such that F.H™ and U.H°° are opposite and, for A. # 1, F.H™ and U.+lH™ are opposite. The spectral numbers (a\,..., aM) at 0 and (c?j,..., c?M) at oo of such an extension С are related by a,- = — Proof. Let U. be a filtration as in (ii). The following explicit construction of a basis of global sections of С from U. is a key idea in [SM3, §3] (cf. also [He4, chapter 5]). The space H°° decomposes into H°° = ™ П Fp П U-p HFpn ?/,_„. G.40) p x^i p These subspaces lead to distinguished spaces of elementary sections g«+p :=zP^a (ЯД,(„ П Fp П f/@ or ,)_„) С Са+Р G.41) for — 1 < a < 0, p € Z, with the properties C" = 0Z-pGa+p Э Gr^^o = фг'б^' = Ga ®zGr»rlCo, G.42) p<0 NGa = (ZVa2 - a)Ga С G.43) p>0 Formula G.43) follows from the monodromy invariance of U. and from z Vg. — a — — ^r on C" G.8). The purpose of the filtration U. is really the splitting in G.42) of the filtration of С by Grf,z'?o- One can choose a basis of elementary sections s{,..., s^ of фа G™ with orders <x(sj) = a,-. They form a basis of Hz for each z € C*. They are principal parts of germs in Co- The point now is that there exist unique germs u,- in Co
124 Connections over the punctured plane -•a+l a a + l Figure 7.2 whose principal parts are s,- and whose higher elementary parts are contained *n ©0>a, Фр>о Z~PGP+P for each i = 1,..., /n: starting with some sections whose principal parts are j,-, one can inductively for increasing i and increasing orders erase all higher elementary parts in 0^>a. @p<qZ~pG^+p by adding elements of Co. The uniqueness of the section e CoП U V P>o G.44) is clear because the difference of two such sections would have an impossible principal part. Figure 7.2 shall illustrate the construction. The eigenspaces Ca should be imagined as columns above the as; the picture does not take into account the different dimensions of the С and the discreteness of the values a. A space G& is nonzero if and only if /J 6 {a!, ..., aM}. Therefore the sections Vi are sums of finitely many elementary parts, so they are global sections in TC(C). They form a basis of Hz for each z 6 C* because the s,- do and G.44) gives a triangular coefficient matrix. The lattice at oo of the Opi -module С := ф; Opt ¦ и,- is ?°o = фС{г) • Vi = 0С|г)л = 0C{?)Ga. G.45) i i a Formula G.43) shows that it is saturated. Going through G.24) and Lemma 7.6 (b) one checks that the filtration on H°° which corresponds to C^ via G.30) is precisely U.. Here one needs zfe = -zdz and Ca — C~a G.37). Obviously, the spectral numbers C?!,..., aM) at infinity are a,- = —aM+i_,-. It remains to go from (i) to (ii). Let ? be as in (i). One can choose a basis vi,..., i>M of global sections in ДР1) whose principal parts ^i,...,sM represent a basis of the space 7.4 Riemann-Hilbert-Birkhoffproblem 125 Co and have orders a,. Then these principal parts j|,..., $д also form a global basis ofthe bundle Я -+ С*, because theyformaClzHz]- basis of V>~°° and because they are elementary sections. They generate а С (?)-lattice C'^ at infinity whose spectral numbers are —a\, • • •, —oCf,.. But the lattice Сж is saturated and contains all elementary parts of the u, (Lemma 7.6 (a)). Therefore C'^ С ?оо- If ?<» were bigger than C'^, the spectral numbers a\,..., aM of C^would satisfy ^a,- + Ylai < 0- But Theorem 7.20 will show Therefore C'^ = Cx. Now one defines ¦Si CC01. G.46) G.47) The identity G.42) holds. The identity G.43) holds because C^ is saturated by assumption. Reading G.41) and G.40) backwards one obtains a monodromy invariant increasing exhaustive filtration U. on H°° which satisfies the prop- properties in (ii) and which is just the filtration corresponding to С,ж via G.30). ? In the singularity case one is interested in C{z}-lattices ?oCP which also are C{{3~' }}-lattices. Then a stronger result holds. Theorem 7.17 Let С be an Ос-free extension over 0 ofthe sheaf H of sections of the flat vector bundle H —у С* with Но С V>~1, and suppose that Co also is a C[{d-1}}-lattice. The filiations which ?0 induces on H°° by G.25) and G.28) are denoted by F. and F?lg. (a) The two conditions for a monodromy invariant increasing exhaustive filtration U. on H°° are equivalent: (i) The filiations F. and U. are opposite in H^°; the filiations F. and U.+\ are opposite in H^forX ф 1. (ii) The filiations F?s and U. are opposite in H™; the filiations F°'« and U.+\ are opposite in H%° for Хф\. (b) There is a one-to-one correspondence between the two sets of data: (i) Extensions of С to Opt -free modules С which have a logarithmic pole at oo and which satisfy: the [i-dimensional space ?(P') of global sections is a good Со/У^1 Co-section. (ii) Filtrations as in (a) (i) and (ii).
126 Connections over the punctured plane Then ?(P]) is also a good ?o/zjCo-section; a basis of it is a C{z}- and C{{ 37 '}}- basis of Cq. The spectral numbers at 0 and oo satisfy a,- — — aM+i_,-. Proof, (a) We will prove that condition (i) for Hf° implies condition (ii) for Hf°. Everything else is analogous. We have Ff'f = F_, =0 because of ?0 С V>~1. For any p > 0 the automorphism<pp := Пы^Ш +k">of Я1°° in <7-29) maPs fph\° t0 ^'*#Г and respects U.Hf°. Therefore on Я,°° <pp(Fp П f/_p) с Fp П [/_„, G.48) (F, П U.p) © С/-Р-, = U-p = <pp(Fp П СЛ.„) Ф C/-P-i, G.49) Я,°° = 0^а>П [/_„). G-50) p The equality dim Fp = dim Fp 8 now implies (Bu-4)> G.51) П U.p = П f/_p G.52) (b) First we go from (ii) to (i). Suppose U, satisfies the properties in (a). Theorem 7.16 and its proof yield an extension С and global sections v,- with principal parts s,- and spaces Ga = фа;=а С - st с С. It remains to show :0 = Ga ф Vr'Giy-'jCo- We need the formula n V^№"«n y-« G.53) for — 1 < a < 0, p > 0, ^ e Z. This follows from the monodromy invariance of C/_? and from z= (—?) = ¦ Ca+l fora>-l G.54) with (a + 1 — j^() being invertible. The formulas G.41), G.53), and the definitions of F, and F°lg show Ga+" = (Grav+P?o) П П П G.55) for — 1 < a < 0, p > 0. This formula together with the analogon of G.40) for Fp shows that G.42) is also valid for a > -1 if one replaces z by V^~', С = 0 VlGa+p D v?Ga+p. G.56) p<0 T 7.4 Riemann-Hilbert-Birkhoff problem 127 This implies that ?(P') is a good ?o/V^'? It remains to go from (i) to (ii). Let? be as in(i)and vi,..., fM be a basis of global sections in ?(P') whose principal parts s\,... ,sfi at 0 represent a basis of the space фа Gr" Ln/V~x Gr°"'?o- It is sufficient to show that they also represent a basis of 0a GfyLo/zGfy^Zo. Then one can apply Theorem 7.16. We may choose v\,..., uM such that c*i,..., aM are their orders. Define Ga:= Ji CC". Then Ca = 0 VapGa+') fora > -1. G.57) G.58) The lattice ?00 is saturated and contains with u,- also its principal part 57 (Lemma 7.6 (a)). Because of dz = —l,2dj. then ?»Э 0 0VsP;Ga+"D 0 C". G.59) — l<a<0p>0 —l<a<0 The sum in the middle generates a CfzHattice ?'„, whose spectrum is al- already —ot),..., — aM. Theorem 7.20 implies in the same way as in the proof of Theorem 7.16 that ?<*, = C'^. Now dz = z" ¦ (—z'S?) again shows Z^ n C° = Ga ф (zZoo П Ca) for a > -1 G.60) and then z^a t 1 -\ —p/-a — v 1J pel, Gr»?o = G"®zGt<v- Now one can apply Theorem 7.16. +p fora lC0 fora >-l, > -1. G.61) G.62) ? Remarks 7.18 (a) The main result Theorem 3.6 in [SM3, §3] can be stated as follows: Under the same assumptions as in Theorem 7.17 there is a one-to-one correspondence between filtrations [/, as in Theorem 7.17 (a) (ii) and good ' W С Со with Z W С W G.63) Comparing the proofs of Theorem 7.17 and [SM3, Theorem 3.6], one finds that these /г-dimensional spaces W are precisely the spaces ?(P') of global sections of extensions С as in Theorem 7.17 (b) (i). But this is quite nontrivial.
128 Connections over the punctured plane M. Saito does not consider the questions of whether a basis of germs in W is extendable to a basis of global sections in the bundle H -*¦ C* and whether they give a logarithmic pole at oo. He does not formulate Theorems 7.17 and 7.16 and does not use the filtration F.. (b) The main application of Theorem 7.17 and of [SM3, Theorem 3.6] is the same: to find a distinguished basis of sections in the GauB-Manin connection of a versal unfolding of a singularity which induces a Frobenius manifold structure on the base. In section 11.1 we will use Theorem 7.17 and Theorem 8.7 in order to extend the GauB-Manin connection in an explicit way to oo. M. Saito uses G.63) in order to apply a result of Malgrange on deformations of microdifferential systems ([Mal3, §5], [Mal5]). In the proof of Malgrange's result there also is an extension of a connection along oo, but other ingredients are a Fourier transformation and microlocal aspects. (c) One can separate the proof of [SM3, Theorem 3.6] into four pieces: (i) to construct from a filtration U. as in Theorem 7.17 (a) (ii) a good ?o/V^'?0-section W С ?o ([SM3, Proposition 3.4], similar to the first part in the proof of Theorem 7.16 with V3"' instead of z), (ii) to show that this section satisfies G.63) (a discussion of principal parts, this is not so difficult), (Hi) to recover a filtration U. from a good ?o/V;j"'^-section W с ?o with G.63) ([SM3, Proposition 3.5]), (iv) to show that this section is the unique section constructed from U. in (i) (this is the most difficult piece; it is not even clear a priori that the elements of a section W with G.63) consist of finitely many elementary sections). (d) We will meet in the singularity case an extension of G.63) to the GauB- Manin connection of a versal unfolding, cf. chapter 11, A1.20) and A1.54). The coefficients will carry precious structure. 7.5 Spectral numbers globally The results in this section are inspired by and partly due to [Sab4, IV 1.10]. We stay in the situation of section 7.1. The spectral numbers (a i,..., aM) of a C{z}-lattice ?0 С V>'°° had been defined by tt(j | a,- = a) = d(a) := dim Gr^?0 - dimGr?"'?0, G.64) cf. G.21) and G.23), and were ordered by ai < ... < aM. The V-filtration is independent of the coordinate z in С (Lemma 7.7), so the spectral numbers of a lattice are also coordinate independent. But in G.9) a total 7.5 Spectral numbers globally 129 order < had been chosen. A different order may give a different V-filtration and other spectral numbers. An example due to M. Saito [SM3, D.4)] will be presented in Remark 11.7. The following result shows that the sum ?, a,- is even independent of the order <. The determinant bundle det H = Ялд —> С* is a line bundle and carries an induced flat connection. The monodromy is given by its eigenvalue det h e C*. The determinant sheaf ?лм of an Cc-free extension ? to 0 of the sheaf H of sections of H -*¦ C* is an Oc-free extension of the sheaf HAIM of sections of Ялм -> С*. If ?0 С V*'00 then the sections of ?лд also are of bounded growth at 0. More precisely, one has the following. Theorem 7.19 Let С be as above with ?o С V>~°° and with spectral numbers a\,...,a^caO. Then the germ C^ of C^ is a saturated lattice in 0»НЛДH with residue eigenvalue Yl?=i ai- Proof: One chooses a C{z}-basis a>\,... ,соц, of ?o whose principal parts s!,..., sM represent a basis of @a Gr^ ?0/zGry~' ?0 and have orders a i,..., ад. Then ft>i,..., а>ц generates ?q^ over C{z). It is sufficient to show a>\ Л ... Л (Op = u(z)si Л ... G.65) with u(z) e C{z), "@) = 1.ThenC^ = C{z}«i л ... Aa>,l= C{z}si л ... л Sfj. has residue eigenvalue ^ or,- (in the rank 1 case a lattice with sections of bounded growth is automatically saturated). The principal parts sit..., 5^ form a CMIz'^-basis of V>~°°, so with unique af) e C, which satisfy au = 1 and af? ф 0 only for 0", k) = (i, 0) or a, < aj + k. G.67) Expanding a>\ Л ... Л ш^ with G.66) one has to see that all combinations of summands except for the combinations of the principal parts are contained in zC{z}s\ A... a s^. This is elementary; it follows from G.67) and the properties in G.9) of the transitive order <. ? Theorem 7.20 Let the following be given: a holomorphic vector bundle L -*¦ M of rank jx on a compact Riemann surface M with sheaf С of holomorphic sections, afinite setT, С М, aflatconnectiononL\M--? such that the sections of ? have bounded growth everywhere. Let {a \ (q),..., <Хц(д)) denote the spectral
130 Connections over the punctured plane numbers of the lattice Cq for q e T,. The line bundle L^ is the determinant bundle ofL. Then = -deg I* G.68) Proof. The restriction L^\M_-L is equipped with the induced flat connection. The sheaf ?лд has logarithmic poles with residue eigenvalues ?,. at{q) at all points q <=Y,, because of Theorem 7.19. One can choose any nonvanishing meromorphic section on LA/X with zeros and poles in a set ?' with ?' П ? = 0. It yields a meromorphic connection 1 -form w with simple poles in X/ U ?. The sum of its residues vanishes. . G.69) П Chapter 8 Meromorphic connections Section 8.1 is a reminder of logarithmic vector fields and differential forms and some other classical facts. In sections 8.2-8.4 extensions with logarithmic poles along a divisor V С M of the sheaf of holomorphic sections of a flat vector bundle on M — V are discussed. In the case of a smooth divisor V in section 3.2, there are three important tools for working with such extensions: the correspondence to certain filtrations, the classical residue endomorphism along V, and the (less familiar) residual connections along V, whose definitions require the choice of a transversal coordinate. Extensions to singular divisors V are treated in section 8.3 in greater gener- generality than in the literature. If an (automatically locally free) extension to Vreg with logarithmic pole is given, then there exists a unique maximal coherent extension to V. It is locally free only under special circumstances. The case of a normal crossing divisor is discussed in section 8.3, the GauB-Manin connec- connections for singularities provide other very instructive examples (Theorem 10.3, Theorem 10.7 (b)). In Section 8.4 only some remarks on regular singularities are made. 8.1 Logarithmic vector fields and differential forms For the reader's convenience we put together some definitions and results from [SK4][Del][Ser], which will be useful in sections 8.2 and 8.3. Let M be an m-dimensional complex manifold, V С М a hypersurface, and g : M —> С a holomorphic function such that V = g~'@) and the ideal sheaf (g) С Ом is everywhere reduced. A (locally defined) holomorphic vector field X e TM is logarithmic if X g e (g). A (locally defined) meromorphic g-form со with poles at most on V is log- logarithmic if ga> and g dtw are holomorphic g-forms. These definitions are obvi- obviously independent of the choice of the defining function g for V. The sheaves of 131
132 Meromorphic connections all logarithmic vector fields and differential <7-forms are denoted Der^ (log P) and ?2^(logP), respectively. Example 8.1 (Smooth divisor P.) Suppose M С Сш and g = zi- Then these sheaves are free CV-modules, ^©ffio^-dz,. Z\ i=2 = /\SlM(\ogV). The logarithmic vector fields are the vector fields tangent to P. (8.2) (8.3) (8.4) This case of a smooth divisor P is essential for the general case because of some codimension 2 arguments. The singular part Vsing С М of the divisor P has codimension > 2 in M. Let i : M - Ря„г «^ М denote the inclu- inclusion. Then i*OM-vsl4 = Ом by the Riemann extension theorem, and also i*((g)\M-v^) = (g). The above definitions show I = *-*(DerM(logP)|M_psiJ, (8.5) i = i*(uqM(\ogV)\M-Vting). (8.6) So the logarithmic vector fields are the vector fields tangent to Preg. Formula (8.4) also holds for general M and P. Example 8.2 ([Del, II §3], the normal crossing case.) Suppose M = Cm and g = z\ ¦ Zi ¦. ..-Zk f°r some к with 1 < к < m,soP = (J*_, С' x{0}xCm~'. Then the sheaves Der^(logP) and S^(logP) are free OM-modules, Э Zi3z7 (8.7) (8.8) i=k+l and the sheaves S^(logP) for 9 > 1 are determined by (8.4). Here equality for the restrictions to M - Vsing follows from Example 8.1. The sheaves on the right are free Ow-modules. Then equality follows from (8.5), (8.6) and 8.1 Logarithmic vector fields and differential forms 133 The sheaves of logarithmic vector fields and differential forms are coherent. This is clear for DerM(logP) because that is the kernel of the map TM ->¦ Ом Kg)* Xh Xg. But it is not so clear for the differential forms. It will follow from (8.9), but first we will see it in a way which will also be useful in section 8.3. The proof uses the following result of Serre. Theorem 8.3 [Ser, Theorem 1 and Proposition 7] Let X be a normal variety, Y С X a subvariety of codimension > 2, and T an Ox-y-coherent sheaf on X — Y without torsion. Let i : X - Y ^ X denote the inclusion. (a) Then i*T is Ox-coherent if and only if there exists some Ox-coherent extension ofT. (b) (i) lfT is reflexive and i*T is Ox-coherent then i*T is reflexive and it is the only reflexive extension ofT. (ii) Especially, if there exists some locally free extension ofT then i+T is this locally free extension. A sheaf is reflexive if the map to its bidual is an isomorphism. The statement in Theorem 8.3 (b) (ii) is elementary and was already used in Example 8.2. Arguments of the following type can be found in [Del, II 5.7] and [Mal7]. They will again be used in section 8.3. Theorem 8.4 Let M be a manifold and P a hypersurface. The sheaves of logarithmic vector fields and differential forms are coherent and reflexive Ом-modules. Proof. Following Hironaka, there exists a resolution / : (M, P) —*¦ (M, P), that is, a manifold M, a normal crossing divisor P С М, and a proper holomor- phicmap/: M ->¦ M with F~\V) = V and such that /: M-P-> M-V and / : M - f'^H M- Vsing are biholomorphic. The sheaf Der^(log D) on M is locally free because of Example 8.2. By Grauert's coherence theorem for proper maps (cf. for example [Fi, 1.17] the sheaf /*Der^(log P) is OM -coherent. It is a coherent extension to M of Der^ (\ogT>)\M-vsing- By Theorem 8.3 (a) and (8.5), then DerM(logP) is coher- coherent and reflexive. The same applies to the sheaves of logarithmic differential forms. ? Reflexiveness of Derw (log P) and ?2lM (log P) also follows from the fact that these sheaves are dual OM -modules by the inner product DerM(logP) x OM, (X, со) v+ co(X) (8.9)
134 Meromorphic connections ([SK4, A.6) and A.7)]). This duality is a consequence of Example 8.1 and (8.5), (8.6). In general, DerM(log?>) and fi},(log?>) are not necessarily locally free OM- modules. If they are, then PcMis called a free divisor. Several criteria for a divisor being free are given in [SK4]. In the case of singularities or, more generally, massive F-manifolds with Euler fields, the discriminants are always free divisors, often the bifurcation diagrams (section 4.3). 8.2 Logarithmic pole along a smooth divisor Let M, V, g be as in section 8.1, M a complex manifold, V с М ahypersurface, and g : M -» С holomorphic with V = g~'@) and everywhere reduced ideal sheaf(s)c0M. Let H -> M — V be a flat vector bundle with connection V and sheaf H of holomorphic sections. Deligne [Del] gave a precise meaning to the notion regular singular for (meromorphic) equivalence classes of Ow-coherentextensionsofW and showed that there is always precisely one regular singular equivalence class of exten- extensions. We will come back to this in section 8.4 K. Saito [SK2] emphasized that one should look not just at this equivalence class of extensions, but at specific extensions of TL which are usually given naturally. Here and in section 8.3 CV-coherent extensions with logarithmic poles will be discussed; the case of a smooth divisor here, the general case in section 8.3. An important fact and tool is that in the case of a smooth divisor V the correspondence G.30) between extensions with logarithmic poles and some filtrations extends to the higher dimensional case. Two other tools in this case are the residue endomorphism and a family of residual connections. Definition 8.5 Let M, V, g be as above and С be an CV-coherent extension of H to M. The pair (?, V) (or simply ?, if the connection is unambiguous) has a pole of order < r + 1 (or of Poincari rank r) along V for some r e Z>0 if 1 , V : С ->• — SUlog?>)®?., (8.10) gr The pair (?, V) has a logarithmic pole along V if r = 0. Remarks 8.6 (a) We are interested here only in the case r = 0. (b) This definition is standard at least in the case r = 0, C, V arbitrary and in the case r arbitrary, С locally free, V smooth. In the second case it is usually formulated in terms of vector bundles [Sab4][Man2]. 8.2 Logarithmic pole along a smooth divisor 135 (c) The definition is independent of the choice of g. (d) Because of the duality (8.9) between logarithmic vector fields and 1-forms, (?, V) has a logarithmic pole if and only if Vx maps С to itself for any logarithmic vector field X. Let(M, ?>)beisomorphicto(Am,{0}x Д'"-'), where Д = {z eC\\z\ < 1} С С is the unit disc, and let g : M -> С be holomorphic with V = g~] @) and (g) С OM reduced. As above, H -> M — V is a fiat vector bundle of rank \i with connection V and sheaf H of fiat sections; i : M - V -> M is the inclusion. The aim of this section is to make the following statements precise and explain them. Theorem 8.7 (a) Any OM-coherent extension С of ТС to M which has a loga- logarithmic pole along V is a locally free О м -module. (b) There is a one-to-one correspondence between sheaves С on M as in (a) and increasing exhaustive flltrations of H -» M — V by flat subbundles FPH -+ M -V, pel. (c) There is a unique О-^-Ипеаг residue endomorphism Res on the locally free Ov-module (C/g ¦ C)\v- It is independent of g. (d) The function g induces aflat residual connection Vrej'« on (C/gC)\v- The residue endomorphism Res is flat with respect to this residual connection. Two functions g and g with 'g/g e O*M induce the same residual connection if(g/g)\v = constant. Remarks 8.8 (i) Often one considers a priori only extensions of the bundle U _>. м — x> to a vector bundle on M with a logarithmic pole along V. This is justified by (a). But in the case of a nonsmooth divisor V this approach would be too narrow. (ii) Some particular filtrations of H -> M - V are the filtrations which are trivial on each subbundle of generalized eigenspaces of the monodromy, that means, 0=Fp_i С Fp = total subbundle for some p 6 Z. They correspond to the sheaves С as in (a) whose residue endomorphisms satisfy the nonresonance condition: no two eigenvalues of Res differ by a nonzero integer. That these extensions С are uniquely determined by the eigenvalues of their residue endomorphisms is classical ([Del, II §5], [Manl]). (iii) A point of view which together with Theorem 7.9 gives Theorem 8.7(b) is formulated in [Mal4, p. 404] (cf. also [Sab4, III 1.20]): A locally free O^x[Q)- module?'on Дх@)сД" = М which is an extension of the sheaf of sections of Я|д.Х@) with a logarithmic pole along {0} can be extended uniquely to a sheaf С on M as in (a).
136 Meromorphic connections (iv) The definition and the uniqueness of the residue endomorphism Res are classical. Res and V"g are presented together in [Man2, II 2.1] and [Sab4, 0.14.(b)], except for the precise dependence of Ves's on g. (v) Although V№* depends on g and is thus less canonical than Res, it will be extremely useful in the case of bundles on M = P1 x M' where the coordinate z on P1 or i serve as g. First we choose coordinates and set M = Am, V — {0} x Am~l, g = z\ ¦ The dependence on the choices will be discussed later. We have to repeat the construction in section 7.1 of the elementary sections with respect to z\. Again h = hs ¦ hu = hu ¦ hs denotes the monodromy of the bundle H ->• M -V, with semisimple part hs, unipotent part hu, and nilpotent part N ~ loghu. The fibres Hz, z € M — V decompose into generalized eig- enspaces, Hz = фл Ягд with ЯгД = ker(/is — X). These subspaces form the flat subbundles H^ = {JzeM_x> ^г,х wim sheaves of holomorphic sections. Again е:Нх Д"-1 -+M-V, (?, z2, •. •, zm) н+ (e2*1^ ,Z2,. ¦., z») (8.11) is a universal covering, and the space H°° := {pr о A : H x Am~l -> H | A is a global flat section of е*Я} (8.12) is the /^.-dimensional space of global multivalued flat sections. Here pr: e*H -*¦ H is the projection. The monodromy acts on it, H°° = 0X H?° with Ях°° = ker(fts - X). Then for any A 6 H?° and aeC with e~2jria = X there is an elementary section es(A,a):M-V-* H, z i-> ехрBлгга?) exp(-^)A(f, z2, • ¦ •, Zm) (8.13) here e(^, Z2 zm) — z. Formulas analogous to G.7) hold, especially (N \ N --—A, a I = --—es(A, a), (8.14) 2jt( J 2ni , a) = 0 for/>2. (8.15) Again C" denotes the space of all elementary sections es(A, a), A 6 H?°, with a fixed order a. A basis of С is a basis of sections of the subbundle H^. 8.2 Logarithmic pole along a smooth divisor 137 Now let FPH -*¦ M — V, p e Z, be flat subbundles which form an increas- increasing exhaustive filtration. We construct from this filtration an OM -locally free extension LoiH with a logarithmic pole along V in two ways; the first way is more explicit; the independence of choices is easier to see in the second way. 1st way: The filtration F, H induces a monodromy invariant filtration F.Я00. One may choose a basis Aj e FpU)H™jy г = 1, ¦ ¦ •, M, of H°° which splits this filtration and the eigenspace decomposition. Then С := • es(Ah u(J) (8.16) 7=1 = Mj), —1 < Re(aO')) < 0. Obviously these elementary sections form a basis of the bundle H -*¦ M — V. The extension С of H to M has loga- logarithmic pole along V because of (8.14), (8.15) and the monodromy invariance of F.tf00. 2nd way: Let FpH^) denote the sheaf of holomorphic sections of FPH^)- L := 0 E *'* W» n Ou ¦ Ca+P (8.17) X pel for e-2*ia = X, — 1 < Re(a) < 0, and i : M ~ V ^-> M the inclusion. One sees that (8.16) and (8.17) give the same extension ?. The residue endomorphism for this extension L is the Op-linear map Res = ziVa : LjzxL\v (8.18) on the locally free Op-module C/ziQv of rank д. The residual connection VreStZl on ?/z\?\v is the flat connection whose flat sections are generated by the classes in L/ziC\v of the elementary sections es(Aj, a( j) + p( j)) from (8.16). Obviously Res is Vre"' -flat. The eigenvalues of the residue endomorphism are the numbers a(j) + p(j), j = 1,..., /x. Theorem 8.7 follows from the next lemma. Lemma 8.9 (a) The construction in (8.16) and (8.17) gives all Ом-coherent extensions CofH with a logarithmic pole along V. (b) Let a\,..., а^Ье an Ом-basis of an extension L as in (8.16) and (8.17) and Ai(z)^- + X^i>2 ¦Ai(z)dz,- be the connection matrix with respect to it. Then Ai@, Z2, • ¦ ¦, zm) and 5Z/>2 ^'@. Zi, ¦¦¦, zm)dz,- are the matrices for the residue endomorphism Res and the residual connection Vres'*' with respect to the basis g\\v, . ¦ ¦, ctm|d of Ljz\L\v- (c) The residue endomorphism Res is independent of the choice of coordi- coordinates. The residual connection Ves>Zl depends only on z\, not on Z2, ¦ ¦ ¦, zm, andWKS'z> = Ves'?l if(z\/z\)\v = constant.
138 Meromorphic connections (d) The OM-locallyfree extension OM ¦ С с иКщ ofHw depends only on a, not on (zi, ..., zm)- It is the unique extension with logarithmic pole along V and with a as the only eigenvalue of its residue endomorphism. (e) The construction of С in (8.17) from the filtration F.H is independent of the choice of coordinates. It gives a one-to-one correspondence between increasing exhaustive filiations of H -> M -V by flat subbundles FpH -> M - V and OM-coherent extensions ofH with a logarithmic pole along V. Proof, (a) Let ? be an 0M -coherent extension of H to M with a logarithmic pole along V. The induced extension to A x {0} of the sheaf of sections of H \ A, x {0, is ?' := ?/fe? H h zm?)Ux@!- It is a locally free OA xH) -module of rank /x and has a logarithmic pole in {0} С Л х {0}. The 0(AxfO),o)-lattice C'o is saturated. By Theorem 7.9, ?' and ?(, come from an increasing exhaustive filtration of Я|Дх@) by flat subbundles. The flat extensions to M - V of these subbundles are denoted by FpH -> M - V, peZ. They induce by (8.16) and (8.17) an OM-locally free extension ? of H. Choose an OM-b&sis es(Aj, a(j) + p(j)) of ? as in (8.16). Their restrictions to A x {0} form an 0Ax{o}-basis of ?(,. The sections es(Aj, a(j) + p(j)) also form an i*OM-v-basis of iji. Any germ a € ?0 is a unique linear combination . es(Aj< a{ withal e (i*0M-v)o- We claim au) e OMfi: Any derivative of a by j^,..., ¦?- is also contained in ?0 because it has a logarithmic pole along V; its restriction to A x {0} is contained in C'o. So the restriction to A x {0} of any derivative of a(j) by -f-,..., ^- is in ОдХ{0). This shows o-W e 0м,о and ?0 С ?0. Now any germs "ox,..., стд е ?0 with о-;|(дХ{о),о) = «(А,, а(У) + рО'))|(дХ(о),о) generate ?0, because the co- coefficient matrix expressing O\,..., crM by the elementary sections is invertible with holomorphic entries. Thus ?0 = ?o. (b) This holds for any basis of ?0, because it holds for a basis of elementary sections as in (8.16). For such a basis, the connection matrix is A{ ^ with A\ a constant matrix. (c) A change of coordinates does not change A i @, zi, ¦ ¦ ¦, zm) and Res. The definition of the elementary sections depends only on zi and they determine V"-Zl. One sees easily that a change of zi to a coordinate z\ with (zi /zi)\v = constant does not change ?/>2 A,@, z2,..., zm)dz,-. 8.3 Logarithmic pole along any divisor 139 (d) The corresponding statements for the restriction to A x {0} hold because of section 7.3. One can use this and (a). But of course, the uniqueness of the extension of i/H.(X) to an OM -locally free sheaf with logarithmic pole along V and a as the only eigenvalue of the residue endomorphism is classical ([Del, II §5], [Manl], cf. Remark 8.8 (ii)). (e) The first statement follows from (d), the rest with (a). ? 8.3 Logarithmic pole along any divisor Let M, V, g be as in section 8.1, M a complex manifold, V a hypersurface, and g : M -> С holomorphic with V = g~'@)and(g) С ОМ reduced everywhere. Let H -> M - V be a flat vector bundle of rank д with sheaf of holomorphic sections H. Section 8.2 gives a correspondence between certain filtrations and coherent extensions to M — Vsing of H with logarithmic poles along V - Vsing. To make this precise one can choose some Riemannian metric on M and the correspond- corresponding distance function d : M x M -*¦ R>q. For sufficiently small e > 0 and S > 0 with S <K e the submanifold M5,e := {z € M | d(z, V) < S, d(z, Vsing) > e) (8.19) is diffeomorphic to a disc bundle over V - {z € V \ d(z,T>sing)<s}, having the same number of components as Vreg. Theorem 8.7 implies the following. Lemma 8.10 Any OM-v,hs-coherent extension of H to M - Vsing with a logarithmic pole along V - Vsing is locally free. There is a one-to-one corre- correspondence between such extensions and increasing exhaustive flltrations by flat subbundles of the restriction of H to Mg:S — V. The aim of this section is to study extensions to Vsing of sheaves as in Lemma 8.10. The main result is Theorem 8.11. It w ill be proved at the end of the section. Let j : M — Vsing ^Mbe the inclusion. Theorem 8.11 Let С be a locally free extension ofHtoM- Vsing with a logarithmic pole along V — Vsing. Then 7*? is OM-coherent and has a logarithmic pole along V. It is the only reflexive extension of С Example 8.12 Consider M = C2, g = zizz, V = g~\0), H = С х (M-D) with the trivial flat connection. Each ideal (zf', Zj2) for h ,k2 e Z>0 is invariant
140 Meromorphic connections under zi3Zl and цдц- So each such ideal is an Ом-coherent extension of Om\m-v with a logarithmic pole along V. All these extensions coincide on M — VSing. The only reflexive and even locally free extension under them is OM. The last two statements illustrate Theorem 8.3 (b). Example 8.13 Consider M = C3, g = Z1Z2Z3, V = g~'@), H = С2 х (M — V) with the trivial flat connection and global basis e\, e2 of flat sections. The sheaf С := OM ¦ + ?>м + Ом ¦ Z\Zi(e\ + е2) is coherent and has a logarithmic pole along V. The restriction to M — {0} is locally free. The sheaf С satisfies ? = MC\m-v,J (8.20) and is therefore reflexive (Theorem 8.3 (b)), but it is not locally free. Let us consider now the normal crossing case, M = Am, g = zi Zk, V = g"'@) = LM=i DU) = U;-i z7'@); here A = {z e С | \z\ < I). The flat bundle Я -+ M -V has к commuting monodromies Ло) = a?/) -A^, у = 1,... Д, for the standard loops. A universal covering is e : x A m~* M - T>, i...., ft, (8.21) The space of multivalued flat sections H°° is defined as usual, H°° = {pr о A : H* x Am~k -> Я | A is a global flat section of e* Я}. (8.22) All monodromies act on it. The indices of the simultaneous generalized eigen- space decomposition H°° = фл Н?° can simply be considered as tuples X = (XA),..., X(k)) of eigenvalues for the monodromies hA\ ..., A№). The notion of elementary sections from sections 7.1 and 8.2 generalizes: if one chooses A e Я?° and a = (aA),..., a№)) w4ith e-2jria0) = xU) then the natural generalization of the formulas G.5) and (8.13) yields an elementary section es(A,a). Lemma 8.14 will show that any Cw-coherent extension of H with logarith- logarithmic pole along V is generated by elementary sections. It is essentially due to Deligne, but was first formulated in [EV, Appendix C]. We need the following generalization of nitrations on H°°: a Z* -filtration (Pi \^k consists of subspaces 8.3 Logarithmic pole along any divisor 141 Pi С Я00 which are invariant with respect to all monodromies AA),..., h(k) and which satisfy for a suitable m > 0 Pi =0 if lj < -m for some j, (8.23) Pi = H°° iflj > m for all j, (8.24) PL С PL if lj < I) for all j. (8.25) Then Pi = фх Р;,х- A Z^-filtration (Pi) induces к increasing exhaustive monodromy invariant nitrations F.(;) on Я°° by F^ := U P,. (8.26) For X^ an eigenvalue of /гУ) let a^} be defined by e~^iaW = ),<¦» and -1 < Lemma 8.14 [EV, Appendix C] Fix a Zk-filtration (a) For any l_ e ([—m, m] П Ъ)к and any X, choose generators АЦ of Р;,л. The sheaf (8.27) these I X w an Ом-coherent extension ofH with logarithmic pole along V. (b) Fix a Z* -filtration PL and С as in (8.27). The filiations F^ on H°° correspond together to one filtration on Ms,e — D (formula (8.19)j by flat subbundles. The OM-vsin -coherent extension of Л which corresponds to this filtration in Lemma 8.10 is C\M-vsing- (c) Formula (8.27) yields a one-one correspondence between Ък-filtrations and Ом-coherent extensions ofTi with logarithmic pole along V. (d) The sheafX as in formula (8.27) is reflexive if and only if for all I. (8.28) Then С = J*(C\m-vs,,is)- (e) The sheaf С as in formula (8.27) is a locally free OM-module if and only if (8.28) holds and the filtrations F^j) have a common splitting. (f) If С as in formula (8.27) is reflexive then it is locally free outside of the intersection of at least three components D(;). Proof, (a) and (b) follow from the definition of elementary sections, (c) See [EV, Appendix C] for the proof of this.
142 Meromorphic connections (d) The sheaf ? is 0w-coherent, and C\M-vJII:, is locally free. By The- Theorem 8.3 ? is reflexive if and only if ? = Л(?|м_р ). Then it must be generated by the elementary sections which correspond to the Zk -filtration in (8.28). (e) The direction 'if is clear. For 'only if one can compare a generating system of elementary sections as in (8.27) and an 0M,o-basis of ?0. It is not hard to see that one can replace the basis elements by certain elements of the generating set. This 0M,o-basis of elementary sections yields a common splitting of the filtrations F.(j). (f) Any two filtrations admit a common splitting. Q Remarks 8.15 (a) A special case of Lemma 8.14 (e) is the case when the filtrations on the simultaneous generalized eigenspaces H?° are all of the type 0 = Fp«.x.)-i c FpU,x) = Яоо for some numbers p(it k) ? z> When these numbers p(i, k) depend only on i and A<;) then the residue endomorphisms on the components D(i) satisfy the nonresonance condition (Remark 8.8 (ii)): no two eigenvalues of one residue endomorphism differ by a nonzero integer. The Ом-locally free extension of H to M for the case when the real parts of all eigenvalues of all residue endomorphisms are contained in [0, 1) was called the canonical extension by Deligne [Del, II 5.5]. (b) The main arguments in the proofs of Theorem 8.4 and Theorem 8.11 are due to Deligne [Del, II 5.7]. (c) In both proofs one could use [KK, Proposition С1.2 (ii)] instead of Serre's result Theorem 8.3. Proof of Theorem 8.11: One chooses a resolution / : (M, V) ->• (M, V) as in the proof of Theorem 8.4. The first step is to extend the lift f*C to those components of Vrcg which are mapped by / to lower dimensional subsets in M. This can be done by choosing locally around these components for example the filtration 0 = F~l С F° = (whole local flat subbundle) and applying Lemma 8.10. Then Lemma 8.14 (b) and (d) yields an O^-coherent extension to M. Just as in the proof of Theorem 8.4, the direct image sheaf on M is an OM-coherent extension of ? to M by Grauert's proper mapping theorem. Serre's result Theorem 8.3 shows that j*C itself is ©^-coherent and reflexive and the unique reflexive extension of ? to M. It has a logarithmic pole along V because ? has a logarithmic pole along V - Vsing: The sheaf ? and therefore also j*? are invariant with respect to logarithmic vector fields. D 8.4 Remarks on regular singular connections 143 8.4 Remarks on regular singular connections Let M, V, g be as in section 8.1, namely M a complex manifold, V с М a hypersurface, and g : M ->• С holomorphic with V = g~l@) and the ideal sheaf (g) С OM everywhere reduced. Let i : M - V -> M denote the in- inclusion. Let H —>¦ M — T> be a flat vector bundle with sheaf of holomorphic sections H. Deligne denned the notion regular singular not for a single CV-coherent extension of "H, but for equivalence classes of such extensions with respect to a meromorphic equivalence relation ([Del, II2.12]). This equivalence class determines and is determined by a coherent OM[*]- module in j, W which extends H (cf. forexample [Mal6][Mal7]). Here OM [*] = OM[g~'] is the coherent ring sheaf of holomorphic functions onM-D which are meromorphic along V. Deligne showed that there is a unique OM[*]-coherent extension of H whose sections have moderate growth with respect to the multivalued flat sections (for the result and for the notion of moderate growth see [Del, II 5.7, 4.2, 4.1, 2.17,2.10,... ]. One says that this extension is regular singular along V, or the connection is regular singular with respect to this extension. Following K. Saito [SK2, E.1)], one can also say that an Ow-coherent exten- extension of H has a regular singularity along V if it is contained in this meromorphic regular singular extension. In the 1-dimensional case M = Д = {z 6 С | \z\ < 1), 2? = {0}, the regu- regular singular extension is the OM[*]-submodule of i*H with germ V>~00 at 0 ([Del, II §1]). In [Del, II 4.1] several criteria for regular singular are given. OM-coherent extensions with a logarithmic pole along V were used only in the normal cross- crossing case. In view of Theorem 8.11 one has the following. Theorem 8.16 Let ?mewm be an OM [*]-coherentsubsheafofi*Tt which extends TC. The following conditions are equivalent: (i) The sheaf Cmewm is the regular singular extension, that means, its sections have moderate growth with respect to the multivalued flat sections, (ii) The sheaf Cmewm contains one (and then all) Ом -coherent extensions) of "H with a logarithmic pole along V. (in) The restriction of ?mem" to an open set U С М which intersects each component ofVreg contains one (and then all) Ом -coherent extensions) ofH\u-v with a logarithmic pole along V.
144 Memmorphic connections Of course, this is well known, (e.g. [Lo2, 8.10] and references there). It was the starting point for the definition of regular holonomic differential and microdifferential systems ([KK][Bj]). There the meromorphic regular singular extension is called a Deligne-type module (or D-type module). Property (iii) says that regular singular is a codimension 2 property: it is sufficient to check it outside a codimension 2 subset. Chapter 9 Frobenius manifolds and second structure connections The definition and elementary properties of a Frobenius manifold M are put together in section 9.1. Sections 9.2, 9.4, and 9.5 are devoted to their second structure connections. These are connections overP1 x M on the lifted tangent bundle of M with logarithmic poles along certain hypersurfaces. They come from some twists of the original flat structure by the multiplication and the Euler field. To know them is very instructive for the construction of Frobenius manifolds for singularities, because in that case one of them turns out to be isomorphic to an extension of the GauB-Manin connection. Sections 9.2, 9.4, and 9.5 build on the definition and discussion of the sec- second structure connections in [Man2] for the case of semisimple Frobenius manifolds, on results in [Du3], and on [SK9, §5], where they together with many properties had been established much earlier implicitly in the case of singularities. The second structure connections have some counterparts, the first structure connections, which are better known. The latter are partly Fourier duals. The main purpose of their treatment in section 9.3 (and in section 9.4) is to compare them with the second structure connections. 9.1 Definition of Frobenius manifolds Frobenius manifolds were defined by Dubrovin [Dul][Du3]. We follow the notations in Manin's book [Man2, chapters I and II]. Here all manifolds will be complex. In the following, M denotes a manifold with dim M = m > 1 (in the singularity case m = /z). A {k, Z)-tensor is an О и -linear map T : T$k -» T$l. Here TM is the holomorphic tangent sheaf, OM the structure sheaf. A metric g is a symmetric nondegenerate B, 0)-tensor, a multiplication on the holomorphic tangent bundle TM is a commutative (i.e. symmetric) and associative B, 1) tensor. 145
146 Frobenius manifolds and second structure connections The Lie derivative Lie^T of a (k, /)-tensor along a vector field is again a (k, /)-tensor, as well as the covariant derivative Vx T with respect to a connection V on M. Then Vr is a (k + 1, /)-tensor (cf. Remarks 2.7 and 2.13). The Levi- Civita connection V of a metric g is the unique connection which respects the metric, Vg = 0, and is torsion free, VXY — VYX = [X, Y] for local vector fields X, Y. The following definition is a bit more restrictive than in [Man2, chapter I] because of the unit field and the Euler field. Definition 9.1 A Frobenius manifold is a tuple (M, o, e, E, g) where M is a manifold of dimension m > 1 with metric g and multiplication о on the tangent bundle, e is a global unit field and E is another global vector field, subject to the following conditions: A) the metric is multiplication invariant, g(X о У, Z) = g(X, Y о Z), B) (potentiality) the C, l)-tensor Vo is symmetric (here V is the Levi-Civita connection of the metric), C) the metric g is flat, D) the unit field e is flat, Ve = 0, E) the Euler field E satisfies Lie?(o) = 1 • о and Lie^(g) = D ¦ g for some DeC. Remarks 9.2 (a) For (M, o, e, g) as in the definition with condition A) (not necessarily B)-E)) the C, 0)-tensor A defined by A(X, Y, Z) = g(X о Y, Z) is symmetric. It arises from the B, l)-tensor о by contraction with the B, 0)- tensor g. Because of Vg = 0 and the symmetry of g, the potentiality B) is equivalent to the symmetry of the D, 0)-tensor VA. In Theorem 2.15 it is shown that the potentiality is also equivalent to the closedness of the 1-form s := g(e,.) together with the condition Lie*oy(o) = X о Liey(o) + Y о Liex(o) (9.1) for local vector fields X,Y eTM. (b) A manifold (M, o, e) with multiplication о and unit field e and (9.1) is called F-manifold (Definition 2.8 and [HM][Man2]). Formula (9.1) for X = Y = e gives Liee(o) = 0 • o. In this context a vector field E is already called Euler field (of the F-manifold) if it satisfies Lie?(o) = 1 • o. This implies [e,E)=e. (c) The flatness of e, Ve = 0, is equivalent to Lise(g) = 0 and the closedness of the 1-form s = g(e,.) (Lemma 2.16). Because the potentiality implies the closedness of ? (cf. (a)), one could replace D) in the definition by Liee(g) = 0. 9.1 Definition of Frobenius manifolds 147 (d) The potentiality B) written out for arbitrary local fields X, Y, Z is VX(Y о Z) - Y о VX(Z) - VY(X о Z) + X о Vy(Z) - [X, Y] о Z = 0. (9.2) This formula can already be found in K. Saito's papers (e.g. [SK9, C.3.2)]). For flat vector fields it is equivalent to the symmetry of VX(Y о Z) in X, Y, Z and to the symmetry of Xg(YoZ, W)'mX, Y, Z, W (cf. (a)). Therefore it is equivalent to the local existence of a potential F e OM,P with XYZ(F) - g(X о Y, Z) for flat fields X, Y, Z. (e) The property Lie?(g) = D ¦ g written out for local vector fields X, Y is E g(X, Y) - g([E, XI Y) - g(X, [E, Y]) - D g(X, Y) = 0. (9.3) Comparison with Vg = 0 shows that the A, l)-tensor is skewsymmetric, g(V(X), Y) + g(X, VA0) = 0. (9.4) The property Lie? (g) = D -g means that ? is a sum of an infinitesimal rotation, a dilation and a constant shift. Therefore V maps a flat vector field X to a flat vector field V(X) = [X, E] - yX. It is an endomorphism of the local system, in other words, it is a flat A, l)-tensor, VV = 0. The eigenvalues of V? = V + f are called the spectrum of the Frobenius manifold and are denoted by d\,.. ¦, dm. They are symmetric around у and one of them is 1 because of [e, E] = e. We order them such that dt + dm+\-i = D,d\ = 1. (f) If V is semisimple then locally there are flat coordinates h,..., tm such that the flat fields St = ^- are eigenvectors of V with Then (9.5) (9.6) for some г,- е С (g) The multiplication with the Euler field is denoted by U:TM-+TM, X (-> E о X. (9.7)
148 Frobenius manifolds and second structure connections 9.2 Second structure connections Let us fix a Frobenius manifold (M, о, е, E, g) and consider the lift pr*TM —> TM 1 I P1 x M -^ M (9.8) of the tangent bundle to P1 x M. The canonical lifts to pr*TM of the tensors g, o, V, U will be denoted by the same letters. The connection V lifts and extends to a flat connection on pr*TM such that V^Y = 0 for Y e pr~lTM. Here г is a coordinate on С С P' and dz the vector field with Эг z = 1, д,рг-[Ом=0. With the multiplication and the Euler field one can twist this connection on pr*TM in essentially two distinct ways. One obtains two series of flat connec- connections, parameterized by one parameter s e С The first structure connections V^s\ s e C, are meromorphic along {0} x M U {oo} x M. The second structure connections V(s), s 6 С are meromorphic along V U {oo} x M, where V - {(г, t)\U- zid is not invertible on T,M}. (9.9) The first structure connections V(j) are due to Dubrovin [Du3] (Lecture 3). Some of the first and second structure connections are related by some Fourier- Laplace transformations. Here we will concentrate on the second structure connections. We will only give the definition and make some remarks on the first structure connections in section 9.3. Dubrovin considered also V@) and called it the GauB-Manin connection of the Frobenius manifold [Du3, Appendix G]. In fact, we will see in section 11.1 that in the case of a hypersurface singularity f(xo,..., xn) the restriction of V(~ ^ to С x M is isomorphic to the GauB-Manin connection on the cohomology bundle of a semiuniversal unfolding of the singularity and that V@) and V(~^ are isomorphic if n is even and if the intersection form is nondegenerate. The whole series V(J), s e C, was defined by Manin and Merkulov [MaM] [Man2,111.2 and 1.4] for the case of semisimple multiplication. Their definition also works in the general case and is given below in Definition 9.3. But in the case of hypersurface singularities, K. Saito perceived this series of connections much earlier [SK9, §5]: A primitive form for the GauB-Manin connection (i.e. essentially the connection V'~2^) gives rise to a period map, which has not so nice properties in the case of a degenerate intersection form. 9.2 Second structure connections 149 To ameliorate that he (essentially) defined the whole series V(j) and proposed to study the period map corresponding to V@). Theorem 9.4 below is a translation and generalization of parts of the results in [SK9, §5], [Du3, Lecture 3 and Appendix G], [Man2,11§1]. Other properties of the connections V(j> will be discussed in the following sections. Definition 9.3 Denote M := С x M - V. Fix s e С The second structure connection Vw on pr*TM\^ -> M is defined by the following formulas for X, Y e pr *Tm \gi (here consider X as a vector field on M and У as a section in the bundle) = vxy - + i + Theorem 9.4 (a) The connection satisfies for X, Y e pr*Tu\^ bis) (W - zTl(x о у), (W - zT\Y). (9.10) (9.11) is aflat connection on pr*TM\^ and V«((W - г)У) = {U - z)VxY - X о v?>((W - z)Y) = (U- (b) The endomorphism -H(y)- (9.12) (9.13) (9.14) is a homomorphism of flat vector bundles {р It satisfies and especially {рг*ТМ\й, V'5»). (V<jf)*y = As о Д,+1 о ... о Д,+*_,(У) (9.15) (9.16) D-l fork > 1 andY е рг~1(Тм)\м- It is an isomorphism if and only if d-, ^~ s ф Ofor alli=.\,...,m (cf. Remark 9.2 (e)).
150 Fmbenius manifolds and second structure connections (c) The Ofo -bilinear map g ¦ , У) (9.17) is a symmetric and nondegenerate and multiplication invariant pairing on the bundle pr*TM\fa and satisfies giA-^X, У) = -g(X, ASY). (9.18) IfX is 4(~s)-flat and Y is 4{s)-flat then g(X, У) is constant. Hence the connec- connections V(~v) and Vw are dual and g induces an isomorphism {pr*TM\lk, ({рг*ТМ\йТ, (V(J))*), (9.19) here is the canonical induced connection. (d) Considers 6 |Z>o. The bilinear form /(s) on pr*TM\^ which is defined by is (—lJs-symmetric and V^-flat and satisfies /(I+1)(X, Y) = -7«(A,(X , Y) = $(Д_, о ... о А,_2 о Af_,(Jf), Y), (9.20) ), AS(Y)). (9.21) It is called the intersection form of Vw. Of course /@> = g. (e) The restriction of V(i) to {z} x M — V is torsion free for any z 6 C. 7/ге restriction of V@) to {z} x M — Й й f/ге Levi-Civita connection for the restriction ofg to {z} x M —V. Proof, (a) First we prove (9.12) and (9.13). The potentiality (9.2) and LieE(o) = 1 • о give for V-flat fields X and Y о У) - [X,E]oY- ^ - 1 - Л == Lie?(X о У) - Lie?(X) oY+(j-^-s)(XoY) У) 9.2 Second structure connections 151 = X о У + X о Ые?(У) + Xo(j sJ (У). = (W - z)Vx(y) - X о ^V - ^ + s By C^-linearity in X and У one obtains (9.12). Formula (9.13) is a direct consequence of (9.11). For the flatness of V(s) on рг*ТМ\й, it is enough to show and - z)Z) = 0 for flat X,y,Z (9.22) = 0 for flat Х,У. (9.23) Formula (9.22) follows with (9.12) and (9.2) for flat X, У, and (also flat) (V-i+j)(Z)from - ^^((W - z)Z). Formula (9.23) follows from (9.10)-(9.13). (b) The endomoфhism V + \ + s maps V-flat sections in pr*TM to V-flat sections (cf. Remark 9.2 (e)). So it is an endomorphism of the V-flat vector bundle pr*TM with eigenvalues d, - ^ + s,i = 1,..., m on each fibre. The endomorphism (U - z) is invertible on pr*TM\g,. Therefore As is an isomorphism if and only if d; - ^ + * f 0 for all i = 1,..., m. (9.16) follows from (9.15) and the definitions of v?} and As. It remains to show (9.15).
152 Frobenius manifolds and second structure connections We obtain with (9.10) and (9.12) for X € pr*TM\^ and for V-flat Y - z)Y) = ^ + A = AS(V^\(U - z)Y) and in the same manner with (9.11) and (9.13) (9.24) (9.25) The formulas (9.24) and (9.25) extend to arbitrary Y € pr*TM\^, formula (9.15) follows. (c) The pairing g is symmetric and nondegenerate and multiplication in- invariant, because g is symmetric and nondegenerate and multiplication invari- invariant. Formula (9.18) holds because V is an infinitesimal isometry with respect to g. The rest follows from the next two formulas (9.26) and (9.27). Here X,Y,Z € pr*TM\g, are chosen such that Z is any vector field on M, X is a V(~j)-flat section, Y is a V(i)-flat section. . Y)+g((U - z)-\X), VZF) - zrl (z о (v - I - - J - *) (W - гГЧ*), (U - zT\Z о Г) + g(.. Л = 0, (9.26) because V is an infinitesimal isometry with respect to g. Analogously dzg(X,Y)=0. (9.27) 9.2 Second structure connections 153 (d) The endomorphism A_, о ... о As_2 о As-\ maps V(t)-flat sections to V^-flat sections. Therefore lM(X, Y) is constant for V(s)-flat sections X and Y, and /<J) is V(s)-flat, V(i)(/(s>) = 0. It is (-lJj-symmetric because of (9.18). (e) The first statement is obvious from (9.10). The second follows from (c). ? Remarks 9.5 (a) Consider an isolated hypersurface singularity and the base space M of a semiuniversal unfolding. We will see in section 11.1 that some choice (of an opposite filtration to a Hodge filtration and of a generator of a 1 -dimensional space) gives the following: a Frobenius manifold structure on M, an isomorphism from the cohomology bundle with its flat structure to pr* TM \ ^ with V(~5\ The form /(?> corresponds (up to a scalar) to the intersection form on the homology bundle. The map A_| о...оД«_2оДа_1 corresponds (up to the same scalar) to the topological map from the homology bundle to the cohomology bundle, which comes from the intersection form. We have m = ц, and dt = I + a\ — a, and D = 2 — (aM — a\)\ here a\,..., Ыц are the spectral numbers of the singularity (cf. section 10.6. They satisfy — 1 < a\ < ... < ct^ < и and a,- + aM+i_,- = и — 1. The map As is an isomorphism if and only if а, Ф | + s for all i. Therefore A|+t and A_|_1_A: for к е Z>0 are isomorphisms. All the maps A^+k for к € Z are isomorphisms if and only if all а,- ф Z, that means, if and only if the intersection form is nondegenerate. In fact, this also follows from the relation between A_« о ... о А|_2 о A|_i, the form /(i\ and the intersection form. (b) Consider an isolated hypersurface singularity as in (a) with degener- degenerate intersection form. Then the homology bundle has global flat sections: the monodromy group is generated by the Picard-Lefschetz transformations of a distinguished basis of vanishing cycles (cf. [AGV2] for these notions). The elements of the radical of the intersection form in each fibre glue to global flat sections. The cohomology bundle never has global flat nonzero sections. Such a section would give a flat subbundle of rank /x — 1 in the homology bundle such that the quotient bundle (of rank 1) would have trivial monodromy. Because of the Picard-Lefschetz formulas then this subbundle of rank /x - 1 would contain all the vanishing cycles in the distinguished basis. But they generate the homology, a contradiction. This implies V(i> Щ V(-J> and also V(f> Щ V@> Щ V(-5>, because V(J> and V(~5> are dual and V@) is selfdual (Theorem 9.4 (c)). Except for K. Saito [SK9, §5], nobody in singularity theory considered the intermediate connections V(s) for s € (—5, §) П ^Z. Can one describe their monodromy groups using only topological data? Especially for V@)?
154 Frobenius manifolds and second structure connections (c) In the situation of (b), V(f> Щ V@) Щ V(~5> also follows from another result of K. Saito [SK5]: Up to multiplication by scalars, the (degenerate) intersection form is the only flat bilinear form in the homology bundle. This follows essentially from the connectedness of the Coxeter-Dynkin diagram [Ga] [La]. The connection V@) has the nondegenerate flat bilinear form g = /@). (d) One could consider the shift from MtoP'xMas uneconomic because of the following: (i) Because of Liee(o) = 0 and Lie?(g) = 0, the Frobenius manifold is con- constant along e. One could take locally an m — 1 -dimensional slice U transversal to e and equip С x U with the structure of a Frobenius manifold locally isomorphic to M. Because of [e, E] = e the Euler field also extends to С х U. (ii) Suppose M = С x U. One can recover V(j) and g on рг*ТМ\м from the restrictions to {0} x M — T>. They are canonically extended from {0} x M — T> with the flow of Эг + е: With respect to this flow on pr*TM one has Lie derivatives Lie9z+e(U - z) = 0, Liedz+e(g) = 0, and also %%eY = 0 for Y withLieeK = Lie3!r = 0. (iii) In singularity theory, usually the GauB-Manin connection is considered on the base M of a semiuniversal unfolding, and not on a space Л х М for some disc А С С But it is more convenient to have the whole Frobenius manifold as {oo} x M in the base space P1 xMof Vм. One can treat all vector fields in 7д/ in the same way and does not have to distinguish those invariant under Liec. Also, the second and first structure connections are closely related, and for the first structure connections one does not have this possibility to reduce the dimension. 9.3 First structure connections Also the first structure connections are defined on the lift pr*TM (cf. (9.8)) of the tangent bundle of a Frobenius manifold (M, о, е, E, g). The following definition and statements are known and can be found in different versions in [Du3][Man2][Sab4]. Definition 9.6 Denote M = С* х М. Fix s € C. The first structure connection Vw on рг*ТМ\м -> M is defined by the following formulas for X, Y e рг*(Тм)\й (again consider X as a vector field on M and У as a section in the bundle) (9.28) (9.29) 9.3 First structure connections 155 Remark 9.7 If one pulls back the connection V(i) with the involution i : P1 x M -> P1 x M, (z, t) i->- (-z, t), one obtains a connection j*V(i) on pr*TM\fy -*¦ M which is given by the formulas = V3:r + - (v + l- + s) (Y)-Eo Y. + s) (9.30) (9.31) Theorem 9.8 (a) The connection V(i) is flat. (b) The multiplication by z is an isomorphism of flat vector bundles (c) IfX is V(s)-flat and Y is i*V<-~l-s)-flat then g(X, Y) is constant. Hence the connections V^ and ("V*^ are dual. (d) For each z € C* the restrictions to {z} x M of all the connections V(s), seC* coincide. These restrictions are torsion free. Proof, (a) For flat XJ,Ze pr*TM one has V^V^Z = z V%\Y oZ) = z Vx(r oZ) + z2XoYoZ oZ)+z2YoXoZ = here the symmetry Vx(Y ° Z) = Vy(X о Z) follows from the potentiality. ForflatX, Y G pr*TM onehasV^K = 0, V3z(XoY) = 0,andV(r)isflat,so -XoEoY, + z-EoXoY. = XoY These two terms are equal because of the following calculation. It uses LieE(o) = 1 • о and the potentiality (cf. (9.2)) X о Y + VXaYE = X о Y - UeE(X oY) + V?(X о Y) = -Lie?(X) oY-Xo LieE(Y) + VX(E о Y) + [E, X] о Y (b) Formula (9.29) shows that z Y is V(s)-flat if Y is V(j+1)-flat.
156 Frobenius manifolds and second structure connections (c) The endomorphism V is an infinitesimal isometry with respect to g and g is multiplication invariant. (d) This is obvious from (9.28). Q In this paper we will not make use of the first structure connections. Con- Concerning them we restrict ourselves to the following remarks. Remarks 9.9 (a) In Dubrovin's papers the first structure connection V(~5J is a main tool for studying the Frobenius manifolds. Sabbah [Sab2,4. l][Sab4, VII 1.1] approaches and characterizes Frobenius manifolds by the first structure connection Ф(~). More precisely, he considers j* V(t~> where j : P1 xM-> P1 x M is the map (z, t) \-> (—±, t). Also Manin [Man2] considers the first structure connection V(t~>. (b) All the connections V(l) give the same family of flat connections on the submanifolds [z] x M for z e C*. The Euler field is not necessary for their definition. One can regard this family of flat connections as the primary datum and the extensions along 9Z via an Euler field as mere refinements. But the definition of the second structure connections requires the Euler field right from the beginning. (c) Some of the first and second structure connections are related by Fourier- Laplace transformations. One can obtain precise informations following Sabbah [Sab4, V 2]. One has to consider the whole vector bundle pr*TM and global meromorphic sections, in order to make everything algebraic with respect to the variable z. Then one can check that the inverse Fourier-Laplace transformation in the precise sense of [Sab4, V 2.10] gives a correspondence between Г УA~и and V(l) for those s eC such that V+{ + s-kis invertible for all A: 6 Z>j. This is not satisfied precisely for s in (J?Li(~Ф + ^ + Z>i). In the case of a hypersurface singularity with degenerate intersection form (cf. Remarks 9.5) the correspondence is valid for s = — !|, but not for s = |. (d) One can write the defining equations for Vw and V(s) in a different way, emphasizing their similarity. For X, Y € pr~l(TM) С pr*TM one has zXoY, s (9.32) (9.33) (9.34) (9.35) Using (9.13), one can check that (9.34) and (9.35) define Vw. 9.4 From the structure connections to metric and multiplication 157 (e) In the case of a hypersurface singularity f(x0,.. ¦, х„) the second structure connection V(~f' corresponds to the GauB-Manin connection on the cohomol- ogy bundle. In fact, a main point of the construction of a Frobenius manifold for a singularity / is to enrich the GauB-Manin connection to a connection isomorphic to V'"'. Analogously the first structure connection V^J corresponds to a con- connection coming from oscillating integrals. One can also construct a Frobenius manifold for a singularity / via oscillating integrals and the first structure con- connection V(~5-'>. Sabbah generalized this to the case of certain global functions j ¦ у ->¦ С with isolated singularities on affine manifolds Y. We discuss this informally in section 11.4. 9.4 From the structure connections to metric and multiplication The first and second structure connections of a Frobenius manifold (M, o, e, E, g) are defined on the restrictions of pr*TM to the submanifolds M and M of P1 x M. But pr *TM is a canonical extension to P1 x M of their sheaves of sections. One can apply the notions from chapter 8. We will see that the pairs (pr*TM, Vw) and (pr*TM, V(I)) carry most of the structure of the Frobenius manifold. The contents of Lemmas 9.10, 9.13, 9.14, and 9.15 are summarized in figure 9.1 and figure 9.2. The lower half of the first diagram is only proved for a massive Frobenius manifold. First we discuss the metric. The following canonical isomorphisms are im- important in order to shift information from pr*TM to TM. The sheaf prtpr*TM of fibrewise global sections of pr*TM is a free ©^-module of rank in, because pr*TM can be seen as a family of trivial bundles on P1. One has canonical isomorphisms Tm = T{0}*M = 7(оо)хм = pr*pr*TM = рг„рг-хГм. (9.36) Second structure connection V(l) onP'xM log. pole {oo} xM _,,^i^P V7 P1 x {0} log. pole_ multiplication Figure 9.1
irr. P1 X log pole {0} . pole {00} {0} x M x M 158 Frobenius manifolds and second structure connections First structure connection V^ on P1 x M multiplication V+\ + s, V Figure 9.2 Lemma 9.10 [Man2, Ц 2.1.1] (a) The pair (pr*TM, V(s)) has a logarithmic pole along {со} х М. The residue endomorphism and the residual connection with respect to ± are defined on T{oo]xM. Under the identification T{oo]xM = TM the residue endomorphism is V + \ + s, and the residual connection is the flat Levi-Civita connection V of g on M. (b) The pair (pr*TM, V(j)) has a logarithmic pole along {0} x M. Under the identification Тщ^м = TM the residue endomorphism isV + \ + s, and the residual connection with respect to z is the flat connection V on M. Proof, (a) One can use Lemma 8.9 (b). One has to write out (9.10) and (9.11) for a basis of global V-flat sections of pr*TM, using the coordinate z = - and (b) Similarly without z. ? Using g one can read off not only the flat connection V, but also the metric g from the logarithmic poles along {oo} x M of the pairs (pr*TM, Vw). In fact, Lemma 9.11 (a) shows that one can read it off also from the poles along V. Lemma 9.11 (a) and (b) are not profound, but they will be very informative when we come to our version (Theorem 10.13) of K. Saito's higher residue pairing in the singularity case. As usual, the residue of a meromorphic 1-form со on P1 at a point q 6 P1 is resqco = ~ fy a> where у is a small positively oriented loop around q. Lemma9.11 (a)LetX,Y eTMandX,Y e pupr*TM be the lifts to pr*TM. For anyt 6 M g(X, Y)(t) = r«(oo,o g(X, Y)dz = - J2 res(Mg(X,Y)dz. (9.37) 9.4 From the structure connections to metric and multiplication 159 (b) Consider an s e {%>o ond suppose that A := A_s о ... о Д5_| is an isomorphism (compare Theorem 9.4 (b) and (d)). Then I{s) is nondegenerate and induces another nondegenerate bilinear form /(~4) on pr*TM\^ by /(-i) = /(s) о (Д~' x Д~'). The form /(~s) is V('s)-flat. It satisfies for any X, Y as in (a) g(X, Y) = (- , Y). (9.38) Proof, (a) The first equality follows from the definition of g. The sum of the residues of a meromorphic 1-form on a compact Riemann surface is 0. (b) The form I(~s) is V("j)-flat because of Theorem 9.4 (b) and (d). Formula (9.38) follows with (9.16), (9.20), and the (-l^-symmetry of I(s). D For the first structure connections there is no analogon of g. Theorem 9.8 (c) leads to another interplay between metric and the first structure connections. The following version for V^^ is close to [Sab4, VI 2.b]. A quite different version is in Dubrovin's papers (e.g. [Du4, chapter 3]). We will not subsequently use it. But again, it will be informative to compare K. Saito's higher residue pairings in the singularity case with it. As in Remark, г : P1 xM -> P1 xM denotes the involution (z, t) \-+ (-z, t); the induced involutions on pr*TM and OPixW and other sheaves on P1 x M are all denoted by i*. Lemma 9.12 Fix a point q 6 {0} x M С P1 x M. Let gq be defined by gq : (pr*TM)q x (pr*TM)q -+ OrxM,q It is a nondegenerate and OPi xM-sesquilinear pairing, that is gq{f -X,Y) = f- gq(X, Y) = gq(X, i*f ¦ Y) for f 6 OpiXM,?- It is i-hermitian, that means, gq(Y, X) = i*Cgq(X, Y)), and satisfies for Z e {pr-{TM)q, X, Y e (pr*TM)q Z gq(X, Y) = g,(v^X, Y) + gq(X, V^Y) Эг g,(X, Y) = gq (V^X, Y) - gq {X, V^Y) Proof. Everything follows easily from the definitions. (9-39) (9.40) (9.41) (9.42) (9.43) D
160 Frobenius manifolds and second structure connections Now we come to the multiplication. The multiplication of a Frobenius man- manifold (M, о, е, E, g) is encoded in that pole of the first and second structure connections which does not encode the flat structure. The following is an ex- explicit description for the first structure connection. A more general discussion is given in [Sab4, 0 14.c]. Lemma 9.13 The pair (pr*TM, Vw) has a pole of Poincare rank 1 (in the sense of Definition 8.5) along {oo} x M. Fix coordinates t\,...,tm on M around a point q, the coordinate z= \ on (P1, oo), and fix a basis of pr*T M in a neighbourhood of (oo, q) € P1 x M. The connection matrix of V(s) with respect to this basis takes the form 1=1 df,- г (9.44) where all coefficients are in Cp Under the identification Тм = Tjoo) x м, the matrix —f2o(O, t) encodes U and the matrix ?2,@, t) encodes the multiplication by j^. Proof: All statements follow directly from Definition 9.6. ? For the second structure connections, the multiplication is encoded in the geometry of V and the poles of (pr*TM, ^7(i)) along it. This will be made precise in the next section in the case of a massive Frobenius manifold. Here we give only a weak general statement: If X, Y e рг*рг*Тм then also VXF € pr*pr*TM and ^^Imxm = VxF|(oo)xM. Therefore one can recover from V^F the fibrewise global section VXY and the difference V^Y —VXY and then with (9.10) the product X о Y, if V + \ + s is invertible. 9.5 Massive Frobenius manifolds A Frobenius manifold (M, o, e, E, g) is massive if it is generically semisimple. It is semisimple if locally a basis of vector fields e\,...,em exists with e,- oe, = 5,-ye,-. Then these vectorfields are unique up to renumbering. They are called the idempotent vector fields. Semisimple Frobenius manifolds have been studied thoroughly ([Du3], [Man2], [Hi]). The potentiality, or more precisely the condition (9.1) for an F-manifold implies [e,-, ey] = 0 (Theorem 2.11). Coordinates щ, ...,um withe,- = ^ are unique up to shifts and are called canonical coordinates, following Dubrovin. An Euler field E with LieE(o) = 1 ¦ о takes the form E = X!/=i(Mi + ri)ei 9.5 Massive Frobenius manifolds 161 for some г,- б С (Theorem 2.11). Therefore the eigenvalues of U(= E о ) are locally canonical coordinates. Let us fix a massive Frobenius manifold (M, о, е, Е, g). The bifurcation diagram В С Mis the set of points where some of the eigenvalues of U coincide. It is empty or a hypersurface and it is e-invariant, because of LieeW = id. It contains the caustic К, С M, the set of points where M is not semisimple. A function whose zero set is the caustic will be considered in section 14.1. Another important hypersurface is the discriminant V = {t € M | U is not invertible on T,M]. (9.45) The geometry of the discriminant was discussed in sections 4.1 and 4.3. It is very rich. It is a free divisor with sheaf DerM(log T>) = E о Тм (Theorem 4.9). The tangent hyperplanes to T> at smooth points have only a finite number of limit hyperplanes at the singular points. All tangent hyperplanes and the limit hyperplanes are transversal to e (Remark 4.2 (i)). Consider for a moment a Frobenius manifold of the form M=CxM' with С x {t'} the orbits of e. Then one has (Corollaries 4.5 and 4.6): (i) The projection V -*¦ M' is a branched covering of degree m. (ii) The bifurcation diagram is the set of e-orbits through VSing, the caustic K. is the subset of e-orbits through those points of Vsing where the singularities are more complicated than the transversal intersection of local smooth components of V. (iii) The discriminant determines the multiplication. For any massive Frobenius manifold, without assuming M — С х М', one can recover the multiplication from V С P1 x M, because V and P1 x M are automatically big enough. Lemma 9.14 (a) The discriminants V and V are related by V П {0} x M = {0} x V. The canonical projection V ->¦ {oo} x M (or {z} x Mfor any z e C) is a branched covering of degree m. V is a free divisor with DerCxM(logP) = (U - z)(pr*TM\CxM) е). (9.46) (b) Let U С М — В be a sufficiently small open subset with canonical coordinates u\,..., um with E = ]Г,- м,е,-. Then m ОПР1 x U = [J{(z,u)eCx U \z = Ui} i=i consists of m smooth components which do not intersect and which pro- project isomorphically to {oo} x V. The standard lifts 7\,..., ?„, to P1 x U of
162 Frvbenius manifolds and second structure connections e\,...,em are uniquely determined by the following conditions: (i) Their sum is ^e'i = e'(the lift of e), (ii) Each 2J is tangent to all components {(z, u) & <C x U \ z = и j} of V C\ P1 x U with j ф i. This determines the e, and the multiplication on M. Proof, (a) One has to regard the definitions of T> and V and needs Theorem 4.9. Compare also Remark 9.5 (d). (b) The intersection V П P' x U decomposes as described because U С М — B. The conditions (i) and (ii) are obviously satisfied, (ii) because of 2) (z — Uj) = Oforj #i. On the other hand, 2J is not tangential to {(z, и) е С x U \ z = «;} because of e'iiz — ui) = — 1. Therefore e"\,.. .,e"m are uniquely determined by (i) and (ii) and by the fact that they are in рг*рг*Тм. D Lemma 9.14 (b) characterizes the lifts to pr*TM of e\,... ,em as vector fields on P1 x M by their relation to V. Remark 9.16 (d) will characterize them as sections of pr*T M by their relation to V(l). In the case of a massive Frobenius manifold Lemma 9.10 (a) is supplemented by the following. Lemma 9.15 [Man2, II 2.1.1] Let (M, о, е, E, g) be a massive Frobenius manifold. (a) The pair (рг*Тм, V^') has a logarithmic pole along T>. The residue en- endomorphism along any smooth piece of V has eigenvalues (—E +s), 0,..., 0). For s ф — | it is semisimple. For s = — | it is 0 or nilpotent with one 2x2 Jordan block. (b) The monodromy of V*s^ around a smooth piece of T> is semisimple with eigenvalues (—e2*", 1,..., I) for s ? \ + Z. For s e | + Z й is the identity or unipotent with one 2x2 Jordan block. "* Pwof. Equation (9.46) together with (9.10) and (9.11) show that pr*TM is invariant with respect to V(s) for any logarithmic vector field. By Definition 8.5 and (8.9), then (pr*TM, V(i)) has a logarithmic pole along V. For the residue endomorphism we have to be more explicit Let U С M — B, hi, ..., um, E — J^uiei, ЙПР1 x(/, and ё\, ...,em be as in Lemma 9.14 (b). In a neighbourhood of one component {(z, u) \ z — и,- = 0} of V Л P1 x U (z - щ)Ъ?% = -(v+\+s) ^^ej. (9.47) V ^ / z — uj For j ф i locally this is contained in (z — M,)pr*7^f. Because V is 9.5 Massive Frobenius manifolds 163 skewsymmetric, V(?J) 6 Yljфi - Therefore the eigenvalues of the residue endomorphism are (—(| + s), 0,..., 0). The remainder of (a) and (b) follows from Example 7.11 and section 8.2. ? Remarks 9.16 (a) In the case of 5 € \ + Z, it is not easy to say when the local monodromy of V(l) around a smooth piece of V has a 2 x 2 Jordan block and when it is the identitiy. In the situation in the proof it has a 2 x 2 Jordan block for 5 = — 5 if and only if V(e,) ф 0. This is often satisfied. For example, if the discriminant is irreducible one has essentially only one local monodromy and it has a 2 x 2 Jordan block if and only if V ф 0. But in the case of the trivial Frobenius manifold A™, that is, e,- = —, eioej=SiJ, ?= OU g{elt (9.48) one has V = 0, and the discriminant is {u \ u\ ¦.. .-um = OJ.For.? 6 — \ it is even more difficult to give precise conditions. (b) For 5 ф 5 + Z the locally free OcxAf-module рг*ТмIcxM is the unique coherent and reflexive extension of pr*TM\gf\o<Cx M with logarithmic poles along T> such that all residue endomorphisms along smooth pieces of T> have eigenvalues (—(j + s), 0,..., 0). This follows from chapter 8. (c) The situation in the proof has been studied extensively in [Man2, II §3] for V@). One needs to know only the restriction of (pr*TM, V@), g) to one slice P1 x (m@)}. Then one can recover the whole structure because (рг*Тм, V<0)) is the unique extension to a flat connection with logarithmic poles along (V П P1 x U) U {со} x M. The metric g is the V@)-flat extension of its restriction to P1 x {m@)} - V. The data for the slice P1 x {м@)} are called special initial conditions. They completely determine the Frobenius manifold locally. In order to recover the metric g one needs Lemma 9.11 (a). (d) In the situation in the proof, the coordinate z — ui for the component ((z, u) I z — ui = 0} of V П P1 x U induces spaces C"^ of elementary sections ofordera for Vwnearthiscomponent,as in(8.13).Herea e ZU— (|+.s)+Z. The sections ci,..., e~m decompose locally uniquely into a sum (in gen- general infinite) of elementary sections, their elementary parts. For j ^ i + Z the fibrewise global sections 7\,... ,e"m satisfy the following and are uniquely determined by it: they do not have elementary parts in C^ for a 6 —(| + s) + Z<0 and any i. The section 2; is the only one of them with a nonvanishing elementary part in C"^0. This follows from the residue endomorphism. There is a similar statement for
164 Frobenius manifolds and second structure connections s € \ + Z, but it is more complicated and less satisfactory because of Remark 9.16 (a). In all examples of massive Frobenius manifolds which I know, the eigenvalues of V are rational numbers. A partial explanation is given by the following application of the second structure connections and of a result of Kashiwara [Kasl]. Theorem 9.17 Let (M, o, <?, E, g) be a massive Frobenius manifold with a point t e M such that all eigenvalues ofU : T,M ->¦ T,M coincide. Then the eigenvalues ofV and thus also the numbers d\,..., dm and D are rational. Proof. Kashiwara calls a constructible sheaf quasiunipotent if for any map from a sufficiently small disc to the base manifold the induced monodromy around 0 G disc is quasiunipotent. He showed that a constructible sheaf is quasiunipotent if its restriction to the complement of a codimension 2 subset is quasiunipotent [Kasl, Theorem 3.1]. We can apply this to V@) onCxM, taking Vsing as the codimension 2 subset. The monodromies around smooth pieces of V are quasiunipotent, so all local monodromies on discs embedded in С х М are quasiunipotent. Ifz@) denotes the single eigenvalue of U : T,M ->• T,M, thenPnP1 x{t} = {(z@), 0}- The monodromy on С - {(z@), t)} is quasiunipotent. Its inverse is the monodromy around {oo} x M with residue endomorphism V (Lemma 9.10 (a)). Its eigenvalues are the numbers dt — у, i = 1,..., m with d\ — 1. D The assumption for t and U is satisfied for example, if the algebra T,M is irreducible, that means, local (Lemma 2.1). Then E\t is a sum of a multiple of e\t and a vector in the maximal ideal of T,M. Chapter 10 GauB-Manin connections for hypersurface singularities Section 10.1 resumes the definition of a semiuniversal unfolding of a singularity and the resulting structure of an F-manifold on its base space M. This was treated in greater detail in section 5.1. The cohomology bundle is discussed from the point of view of chapter 8 in section 10.2. Sections 10.3, 10.4, and 10.6 put together most of the known results on the GauB-Manin connection of (the semiuniversal unfolding of) a singularity. The presentation is as plain as possible. The detailed discussion of the extensions H(k) for к > 0 (Lemma 10.2, Theorem 10.7) is new and leads to a very explicit approach to the microlocal GauB-Manin system (Theorem 10.10). Also instructive and not well known are two alternative descriptions for K. Saito'shigherresidue pairings, both under some restrictions (Theorem 10.13, Theorem 10.28). The second one gives the link to a polarizing form on the cohomology of the Milnor fibre. The definition and general facts from [He4] on polarized mixed Hodge structures are provided in section 10.5. 10.1 Semiuniversal unfoldings and F-manifolds Let / : (C+1, 0) —*¦ (C, 0) be a holomorphic function germ with an isolated singularity at 0. Its Milnor number /л е N is the dimension of the Jacobi algebra 0/7/:=Ос.+..о/(|?. ¦¦•!?)¦ Unfoldings of /, morphisms between them, semiuniversal unfoldings, and the germ of an F-manifold which belongs to / have been discussed in section 5.1. A semiuniversal unfolding of / is a holomorphic function germ F : (C"+1 x C^ 0) -» (C, 0) with F\(Cn+l x {0}, 0) = / and coordinates Ц,..., ^ on (О*, 0) =: (M, 0) such that the reduced Kodaira-Spencer map l A0.1) a|0 : TQM -+ O/Jf, ^- » Г^Ч(С"+1 х {0}, dtj \_dtj is an isomorphism. 165
166 Gaufi-Manin connections for hypersurface singularities The germ (C, 0) С (C+1 x M, 0) of the critical space is defined by the ideal Jf = (f?, • ¦ •, §~). It is smooth. It is the normalization of the discriminant Ф, 0) = <p((C, 0)) С (С х M, 0) where y:(C"+lxM,0)-*(CxM,0), (x,t)^(F(x,t),t). A0.2) One can choose representatives of all these germs with good properties in the following way ([Lo2, 2.D], [AGV2, 10.3.1]). First, s > 0 is chosen such that /-'@) intersects the boundaries dB1'^ of the balls Bn/X = {x e C"+1 | \x\ < s'} transversally for all 0 < s' < s. Then S > 0 is chosen such that all the fibres f~\z) for z € A := B\ С С intersect dB''+l transversally. Finally, we choose в > 0 and define M := B% such that V does not intersect C A) x M and all the fibres <p~\z, t) for (г,()еДхМ intersect 3B"+1 x {t} transversally. Then the space X := F~'(A) n(Be"+1 x M) and the maps F : X -» A and cp : X —> A x M are good representatives with critical space С С Xofcp and discriminant ?> = (p(C) С А х М. The projections ргс,м : С -*¦ M and ?> -*¦ M are finite and flat of degree fi. The Kodaira-Spencer map (ргс,м)*Ос, X A0.3) where X is any lift of X eTMtoX, is welldefined and an isomorphism of free C^-modules of rank /x. It induces a multiplication о on TM. Then M becomes a massive F-manifold (Theorem 5.3). Generic semisimplicity follows from the smoothness of С and corresponds to the fact that for generic t € M the function F, : X П (B"+l x {t}) -> A (Ю.4) has /x A i -singularities. The caustic К. С М is the hypersurface of parameters t for which F, has less than [z singularities. ^ The field e = a~'(l|c) is the unit field. One could choose F = t\ + F{x, t2,..., fM); then one would have e = ^-; but we will not need this in the following. The field E := a~'(F|c) is an Euler field of the F-manifold (M, o, e) (Theorem 5.3). The Kodaira-Spencer map in A0.3) gives canonical isomor- isomorphisms for all t 6 M, •¦ (T;M,o,?|,)= | (-(-) Jacobi algebra of (F,,je), mult., [F,]). A0.5) \xeSing F, The eigenvalues of U : T,M -» T,M, X (->• ? о X, are the critical values of F, for each t € M. They form on M — 1С locally canonical coordinates и i,... ,um. Locally on M - /C the fields e, = ^ are defined and satisfy e,- о ej = S,ye,-. 10.2 Cohomology bundle 167 There are two discriminants, ?> = (detWr'@) CM with A0.6) V = 0>(CnF-'(O))C {0}x M and V = ^(C) = (det(W - г«й?))~'@) С А х М. (Ю.7) The critical space С and the restriction С П F~'@) are smooth. The pro- projections С -» V and С П F"'@) -»¦ X> are generically one-to-one. Therefore they are the normalizations of V and V, and VС А х MandPcM are irreducible hypersurfaces. The function h := detiU-zid) : АхМ ^ С gives an every where reduced equation for V, similarly detW for V. Both discriminants V and V are free divisors (Theorem 4.9), DerM(logX>) = ?oTM = W(TM), (Ю.8) Here we use notations similar to those in section 9.2: the tangent bundle TM is lifted to A x M, n*TM —> TM A x M M, the canonical lifts of the tensors o,U,e to x*TM ai& denoted by the same letters. Then A0.9) follows from A0.8) and the definition of V, (cf. Lemma 9.14, Remark 9.5 (d)(ii)). A careful discussion of the isomorphisms between semiuniversal unfoldings shows the following (Theorem 5.4). Theorem 10.1 The germ ЦМ, 0), о, е, Е) of the F-manifold and the germs (D, 0) С (M, 0), Ф, 0) С (A x M, 0) depend only on f, that means, they are unique up to canonical isomorphism and independent of the choices of s, 8,6 and F. 10.2 Cohomology bundle Let / : (C"+l, 0) -» (C, 0), F : X -*¦ A, <p : X -» A x M and V С А х М be as in section 10.1. We make the additional assumption that n > 1. That excludes only the AM-singularities in one variable (but not their suspensions
-Г . I 168 Gaufi-Manin connections for hypersurface singularities in several variables). Their GauB-Manin connections and Brieskorn lattices are exceptional. One can find a treatment of some Fourier duals of their GauB- Manin connections and also their Frobenius manifolds in [Sab4, Vll 4.b and 5.c]. In this section the cohomology bundle of <p is discussed using chapter 8 and an argument of Varchenko. In the next section things will be compared with the GauB-Manin connection, that is, with sections coming from holomorphic differential forms. Good references for the following facts are [Lo2] and [AGV2] (but, of course, many of the facts are much older, e.g. [Mi]). Afibre<p-'(z,Oof<p : Д"->. AxM, (x, t) i-»- (F(x, t), t) is singular if and only if (z, 0 e V. Each regular fibre is homotopy equivalent to a bouquet of /x n-spheres. The restriction <p : <p~l(A x M -t>) ->¦ A x M -V A0.11) is a locally trivial C°°-bundle. One may call it a generalized Milnor fibration. The restriction to (Д - {0}) x {0} is a Milnor fibration. The cohomology bundle H":= (J A x M-t> A0.12) has rank /x and a canonical flat structure. Its connection is called V. The cohomology bundle and its monodromy group are essentially indepen- independent of the choice of e, S, в in section 10.1: first, the fibres of a representa- representative for some choice of smaller s', 8', в' are deformation retracts of the corre- corresponding larger fibres of <p by [LeR, Lemma 2.2]; second, the representative <p : X -*¦ A x M is excellent in the sense of [Lo2, 2.D], and therefore the monodromy group does not change when one restricts the cohomology bundle Hn to a smaller base Bxs, x Bg, С Л х М. The subbundle of H" of the Z-lattices Hn{<p~\z, t), Z) is invariant un- under the monodromy group. The local monodromy around T>reg is given by a Picard-Lefschetz transformation: for even n it is semisimple with eigenval- eigenvalues (—1, 1,..., 1), for odd n it is unipotent with one 2x2 Jordan block; in both cases the invariant subspace has dimension /x — 1. Let H be the sheaf of holomorphic sections of H" and i : А х М — V -*¦ A x M the inclusion. With chapter 8 in mind, there are distinguished extensions of ft to Д x M. Lemma 10.2 For any к e Z there is a unique extension 7i№> С i*Ti. of Л to Ax M with the properties: it is an О дхм -coherent and reflexive subsheaf 10.2 Cohomology bundle 169 ofi*Ti., and (Hik\ V) has a logarithmic pole along V whose residue endo- endomorphism along Vreg is semisimple with eigenvalues (^ — k, 0, ..., 0) for 2yi - к ф 0 and nilpotent with a2x2 Jordan block for ^ - к = 0. Proof. By Example 7.11 and Theorem 8.7 there is a unique extension of H to a coherent and, in fact, locally free sheaf on Л х М — Vsing with a log- logarithmic pole along VKg and residue endomorphism as above. By Theorem 8.11, the direct image under j* for j : AxM — Vsing -» Л х М is the sheaf П(к). ? Which of these extensions ft(k) are free Одхм-modules? Following an ar- argument of Varchenko one can read this off from the restrictions to Л х {0}. Let (ti, ¦ ¦ •, t,j.) С Одхл* be the ideal which defines Л х {0} and B(k) := (nik)/(ti,.... fM)ft№))lAX@)- A0.13) The sheaf B(k) is an extension of the sheaf of holomorphic sections В of H" | Л* х {0} to Л x {0}. It is a free Од-module of rank /x. Its sections have moderate growth with respect to the multivalued flat sections of H"\A* x {0}, because they are restrictions of the sections of 7iw, and those have moderate growth (cf. section 8.4, [Del]). We can use the notions of sections 7.1 and 7.2. The germ B^ is a C{z} -lattice in V'00 and has /x spectral numbers SpiB1^) = (af\ .... aj>), denned by G.21) and G.23). The following is essentially due to Varchenko ([Val, ch. 2], [AGV2, ch. 12]). It is a relative statement and will be complemented by the next section 10.3. Theorem 10.3 The space H^ is a free O^Mfi-module if and only ifY!t=\ Proof. Consider /x sections ш\,..., «д of H(k) in a neighbourhood of 0 € Ax M. Choose a basis of Hn(<p~l(z, t), Z) for some (z, t) e Л x M - V and extend the vectors to flat multivalued sections 5b ..., 5M of the homology bundle over Л x M — V. Then det2(((a>/, 5,-)),-;) is a univalued holomorphic function on Л x M - V, because det2 A = 1 for any monodromy transforma- transformation matrix A e GL(/x, Z). The function h — d&tQA — zid) gives an everywhere reduced equation for V. The eigenvalues of the residue endomorphism of Giw, V) along Vreg imply that det2(««,, _ „ . tn-\-2k A0.14)
T1 •г 1 ¦' 'S 170 Gaufi-Manin connections for hypersurface singularities for some function g e OAxM_f,tlitg0 = OAxM,0- The following statements are equivalent: (i) The (jl sections co\,..., co^ form a 6>AxM>0-basis of H^ (especially, then Ho is a free Сдхд^о-module). (ii) They form a basis of Ti(k) in (a neighbourhood of 0 in) Д x M - Vsmg. (iii) The function g does not vanish in (a neighbourhood of 0 in) Л x M—Vsing. (iv) It has value g@) ф 0. On the other hand, the definition of the spectral numbers af},..., а?> in G.21) and G.23) shows that for some function g € 0AiO and that there exist sections co\,..., <ид with Finally, h\A = (-z)" because W is nilpotent on T0M. Therefore there exist sections cot,..., <ыд with g@) ^ 0 if and only if 2 ? af} = /x(n - 1 - 2?). D 10.3 GauB-Manin connection Let /, F, <p: X ^ Ax M, V, and tf" be as in sections 10.1 and 10.2 with n > 1. Holomorphic differential forms yield holomorphic sections in H". The investigation of these sections and their relation to the flat connection on H" is summarized in the notion of the GauB-Manin connection for (p. The GauB-Manin connection for / was introduced by Brieskorn [Bri2]. It has been generalized to complete intersections (as <p) by Greuel [Gre] and K. Saito [SK2]. The GauB-Manin connection for singularities has been studied and applied further by many people. [Mall], [Phi], [Val], [SK6][SK9], [Lo2], [Od2], [SM3], [AGV2], [Hel], and [Ku] are some references with the character of a partial survey. The languages are very different, from an explicit use of integrals to a so- sophisticated use of ©-modules and ?"-modules. Many of the results which we have to cite have been proved several times and in different styles. We will try to present them in the most explicit way and point to more technical ways in remarks. Remark 10.4 Usually the GauB-Manin connection for the semiuniversal un- unfolding of / is developed for the fibration F~'@) -»• M. One can choose F = h + Fi(x, t2,..., t,,,). Then t\ takes the role of z. 10.3 Gaufi-Manin connection 171 In fact, the cohomology bundle Hn on Л х М — V with its flat structure is the trivial extension along dz + e of its restriction to {0} x M — V. We take the uneconomic version with base Л х М because then one can treat all coordinates in M equally without an uncanonical choice of f j. Also it fits better to the second structure connections of Frobenius manifolds (cf. Remark 9.5 (d)). Let r\ € Щ, 0 and со e fi'j^}- ВУ diminishing е,8,в from section 10.1, one can arrange that r\ and со are defined on X. The restriction of ц to a smooth fibre <p~l (z, t) of (p: X -*¦ A x M is closed and can be integrated over cycles in the homology. One obtains a holomorphic section in H". The germ in 0'*7iH is independent of any choice of e, 8, в. So there is a welldefined map fi*,o -»• (J'*^)o- A0.15) The form со induces a relative и-form, the Gelfand-Leray form, which gives on each smooth fibre cp~\z, t) a holomorphic и-form $-\<p~l(z, t). It is the Poincare residue of the form j^j|B"+1 x {t}. Again one obtains a holomorphic section in H". Again this gives a welldefined map n+l A0.16) This also all works for the restriction to A x {0}, that is, for the Milnor fibration of / instead of (p. The sheaf of holomorphic sections of the restriction of H" to Д* x {0} is called В as in section 10.2. Let i0 : A* -> Л be the inclusion. Then analogously to A0.15) and A0.16) one has welldeflned maps C" c+'.o (io*B)o, ¦ О'о*Д)о> A0.17) A0.18) and the images are called Щ and Hq. They are the restrictions of the images of A0.15) and A0.16) to (Л х {0}, 0). An elementary and classical, but crucial task is the analysis of A0.17) and A0.18) for the case of the A,-singularity / = x$ + ... + x2n. Let <S(z), z 6 A* denote a A or 2)-valued family of representatives of the vanishing cycles in the fibres /-'(z). Then one finds (e.g. [AGV2, p. 294]) Г dx0... dxn JS(z) Г Л(г) xod*! ...dxn = cy ¦ z~ d/ -CfZ 2 A0.19) /"'(г) for some с i, c2 6 C*. Therefore the images in A0.17) and A0.18) are V! V "r where V is defined as in section 7.1. and
¦f J 172 Gaufi-Manin connections for hypersurface singularities If / is any singularity with F and <p as above then the fibres cp x(z, t) for (z, t) 6 f)reg have only A \ -singularities. The discussion of the A i -singularity and the definition of 7i№) in Lemma 10.2 imply that the images of the maps A0.15) and A0.16) are contained in 7i^l) and Tif* and that H'Q С В?~° and Щ С 0Q *• We w'^ not Prove the following theorem, but comment upon the difficult and simple parts of it and several ways to prove it. Theorem 10.5 (i) There are equalities H? = B^andH^ = в?~°. The spectral numbers (af\ ..., a^0)) =: (au ¦ ¦ ¦,a^) =: Sp(f) of Щ are contained in (— 1, n) П Q and satisfy a,- + aM+i_; = n — I, the spectral numbers of #q are (ii) The spaces H^ and H{q1) are the images of the maps A0.15) and A0.16). They are free OAxM,o-modules. The sheaves H@) and H(~n are free OAxM- modules. (in) The inclusion Ti(^n С Ti@) is represented by [r]] i-* [dF л tj] for П e Q"Xi0. (iv) The covariant derivatives of sections in 7i(-1) andU^ can be described in terms ofn-forms r\ and n + 1 -forms a> by Vdz[dFAt1]=[dr1], [dF 1 —d/? , [dF 1 Va/3liM = [Lied/atia] - V8 I — со . A0.20) A0.21) A0.22) (v)TheoperatorVszyieldsisomorphismsVgz : H'Q -*¦ Hq andVgz : Ho —>¦ nf\ The space Щ is a free C{{d-l}}-module. Remarks 10.6 (a) The C{z}-lattices Hq and H'a were considered first by Brieskorn [Bri2]. The lattice H^ is called the Brieskorn lattice. Its spectral numbers (a.\,..., a^) form the spectrum Sp(f) of the singularity / [AGV2]. (b) On the one hand, Hq is defined via n + 1-forms. On the other hand, it is determined by the discriminant V С А х М and the monodromy group of the cohomology bundle, because of Щ = B^ and the definition of B>0 and 7i@). The discriminant determines the singularity up to right equivalence (cf. Corollary 4.6, Remarks 5.5 (iv) and (v), [Sche2], and [Wir]). One may ask how much information H^ contains. This leads to Torelli type questions (cf. [Hel]-[He4] and section 12.2). (c) Parts (iii) and (iv) are not deep. Part (iii) follows from the definition of the Gelfand-Leray form, (iv) can be proved with the residue theorem of Leray (cf. for example [Bri2]). 10.3 Gaufi-Manin connection 173 (d) Formula A0.20) shows that V3; : H^ -* Щ and V3: : ft*,"" -» ?^0) are surjective. The injectivity follows from the algebraic descriptions in A0.23)- A0.26) below. The injectivity of V3; : Щ -»¦ Щ also follows from H^ с V>0, which is part of (i) and which was proved by Malgrange [Mall]. Together with Lemma 7.4 this shows that Щ is а С{{Эг"' }}-lattice. The injectivity of Va; : 7i^[) -*¦ H^ also follows from the fact that H'l\Ax{l] for t generic does not have global flat sections, see Remark 9.5 (b). We will come back to it in Theorem 10.7. (e) The relation (a',',..., a?~") = Sp(f) + 1 in (i) follows simply from the isomorphism Vs, : H$ -*¦ Hq. But the symmetry a, + aM+i-; = и — 1 and also — 1 < ot\ < ... < <Хц < п are profound. There are two ways to prove them. One way uses Varchenko's mixed Hodge structure, which comes from Hq, see Remark 10.31. The other way uses K. Saito's higher residue pairings, see Theorem 10.28 (v). (f) Part (i) gives the identity ][]«; = д ^-. In view of Theorem 10.3, it implies (ii) and is equivalent to the freeness of Тц* as an 0дхм,о-пкх1и1е. Varchenko [ Val] [AG V2] gave a proof of the freeness of H^ and thus also of this identity, which is simpler than either of the two ways to prove (i). First, one has to somehow see it for the special singularities x$ H h jc^. Then he shows that any singularity turns up in a semiuniversal unfolding of such a special singularity. Using determinants as in A0.14) he proves that the freeness of Ti^ for x$ + ¦ ¦ • + x^ implies the freeness of Ho' for the other singularity. (g) Greuel [Gre] gave the first proof that the images of the maps A0.15) and A0.16) are free 0A)<M,o-modules. In view of the discussion before Theorem 10.5 this implies that the images are 7ц~]) and Ti0 \ He also determined the kernels (see also [SK6] [SK9] [SM3]). We set S := A x M, S' := A x M - V, X' := (p~l(Sr). Then A0.23) A0.24) A ""' л A0.25) A0.26) The morphism <p : X —v A x M is Stein. This allows cohomology classes to be represented in smooth fibres^)"'(г, f)by fibrewiseglobalholomorphicn-forms.
174 GauS-Manin connections for hypersurface singularities Greuel and K. Saito obtained H@) =S A0.27) 00.28) Coherence and freeness of these extensions of the relative de Rham cohomology for <p : X' —> S' are profound results. But we will not explicitly make use of the precise form of the denominators in A0.23)-A0.26). The next result, Theorem 10.7, shows which of the (Од х м -coherent) sheaves H{k) from Lemma 10.2 are free OA xM -modules and extends part (v) of The- Theorem 10.5 to all of the sheaves H{k). For к < 0 this can be found essen- essentially in [SK6][SK9], but not for к > 0. We give a proof. But for the fact in part (c), that H^ is a free Ом,о{{д~1}}-той\А& of rank ц,, we have to refer to [SM3]. /=o Лч i=0 A0.29) is the ring of microdifferential operators of order < 0 at (dz, 0) б Т*( Д x M) [Phl][SM3]. As in A0.10), 7Г : Д x M -+ M is the projection. The sheaves nJiP of fibrewise global sections are OM-modules, but not coherent 0w-modules, they are too big. Theorem 10.7 (a) The covariant derivative Va.. maps W(t) surjectively to for any tgZ. For any к е Z it yields an isomorphism of Ом-modules and an isomorphism A0.30) A0.31) ffej The sheaf V.^ is a free OAxM-module if and only if it is a locally free OAxM-module and if and only if the germ H^ is a free OAxMfi module. If a, <? Zfor all i then all H^ are free О АхМ ^-modules. Else, K^ is a free OAxM,0-module if and only if к < min(a,- | a, 6 Z). (c) Each germ H^ is a free О Mt0{{d~1}}-module of rank ц. 10.3 Gaufi-Manin connection 175 Proof, (a) From the definition of H(k) it follows that V3; maps H{k) |(Д х М - Vsmg) surjectively to H(k+l)\(& x M - Vsing). Therefore V3; maps H(k) to H<~k+I), but the surjectivity near Vsing still has to be proved. For A0.30) and A0.31), first consider H"\(A x {f} - V) for some generic t. Then Ft has ц. А \ -singularities and ц. critical values. We claim: (i) Any global holomorphic section a in Hn |(Д x {/} — V) is the image under Va. of a unique holomorphic section in H" |(Д x {f} - V). This claim shows that A0.30) and A0.31) are injective. It extends Remark 9.5 (b) that Я'!|(Д x {t) - t>) has no nonzero global flat sections and follows in the same way. Proof of the claim: One chooses a system of nonintersecting paths (e.g. also ordered anticlockwise) in Д x {r} - V from a regular value to the ц, points in Д x {f} П V. One obtains a (distinguished) basis of vanishing cycles. The monodromy around a point (z, t) e Д x {f} П V is given by a Picard-Lefschetz transformation. Therefore the space of preimages of a under Va2 in a neighbourhood of (г, t) in Д x {f} П V forms ад-l dimensional affine space. The difference of two preimages is contained in the д — 1 dimensional vector space of flat sections which vanish on the vanishing cycle that corresponds to (z, t). One extends the д affine spaces of dimension /л — 1 of local preimages of a under Va; along the paths to the regular value. There they intersect in one point because the vanishing cycles form a basis. This shows the claim. 0 Now suppose that a e 7Г*?^№+1), that is, a is a section of Tiik+1) defined in ti~x{U) for some open subset U С М. Let Вм С М be the bifurcation diagram, the set of t e M such that F, has less than /x different critical values. It contains the caustic 1С, the set of parameters t e M such that F, has less than д critical points. We claim: (ii) The unique preimages under V3; in generic slices Д x {r} - V as above glue to a holomorphic section in n~[(U — Вм) — i>. (iii) The holomorphic section is in я»^*\ (iv) It extends to n~\M - /С), (v) It extends to ж ~' (M). (ii)-(iv) are easy to see. But unfortunately, (v) is not so clear a priori. We claim that it follows from
176 Gaufi-Manin connections for hypersurface singularities (vi) The map in A0.31) is an isomorphism in the case of an ^-singularity Proof of (vi) =>¦ (v): It is sufficient to show that the preimage under V3. of the above section a extends to slices Д x {t} — V for generic t e. 1С. Then it is defined outside of a codimension 2 subset of n~l(U) and extends globally. For generic r@) e K, the function F,v» has /x — 2 A \ -singularities and 1 A^- singularity and /x — 1 critical values. One can extend the proof of claim (i) if one can show that a has a preimage under V3_ in a neighbourhood in Д x M of the critical value of the ^-singularity. In such a neighbourhood one can split the cohomology bundle into a subbun- subbundle of rank 2 for the versal unfolding of the A2-singularity and into a subbundle of rank /x — 2 which is invariant with respect to the local monodromies. The subbundle of rank 2 with its fiat connection is induced by the cohomology bun- bundle of a semiuniversal unfolding of the A2-singularity. Therefore (vi) would give locally a preimage of a under Va, and would allow the proof of claim (i) to be applied. 0 The eigenvalues of the monodromy of an ^-singularity are ±i for any number of variables. Therefore (vi) follows from the claim Ak) Ak+\) (vii) The map Vg2 : Tl0 —> Tl0 is an isomorphism in the case of a singu- singularity with a,- $ Z for all i. Proof of claim (vii): When к < —1 it follows from Theorem 10.5 (v) and from claim (iii) (in fact, without any condition on the spectral numbers). When к > 0 one can argue as follows. Let a\,..., aM be an Одхл^о-basis of Tif\ Their restrictions to (Д x @},0) generate the C{z}-lattice Щ' with spectral numbers a\,..., aM. The restrictions to (Д x {0}, 0) of the derivatives Vg+1cr|, ..., Vg+1CTM generate a C{z}-lattice with spectral numbers at\ —к —I, ..., ад — к — 1. Because of claim (iii) and Theorem 10.3 this C{z}-lattice is 4*+1),and?4f+1) is a free 0AxM,o-modulegeneratedby V^+'ct,, ..., v?+l<rM. By induction on k, starting with к = — 1, one can show that V^H^ is an Одхм.о-module. Then V*+1?C = VbiHf = W<f+1). ' 0 (b) The first statement is clear because Д x M is Stein. The second statement has been shown in the proof of claim (vii). Suppose that there exist a,- e Z. Because of (a) we have B(t) = V* Щ for any к € Z (with Bw С V>0 for к < -1). Looking at the derivatives of principal 10.3 Gaufi-Manin connection 177 parts, one sees easily that the spectral numbers a\\ ..., a^ of Bik) satisfy (af\ ...,a™) = Sp(f) -k fork< min(a,- | a,- € Z), A0.32) ^p - *) One can apply Theorem 10.3. min(a, | a, € Z). A0.33) 0 (c) In the notation of [SM3, §2], Hf is the germ (F_n J° OxH of the filtered GauB-Manin system for cp : X -*¦ A x M which is constructed there. By [SM3, §2 E.17) and E.18)] it is a free 0M,o((t' }}-module of rank ц. Because of the isomorphism A0.31), then all germs Hq are free 0м,о{{Эг Х }}-modules of rank /x. This finishes the proof of Theorem 10.7. ? The following considerations on the quotients W(t)/W(* l) are due to K. Saito [SK6][SK9] (for к < 0). Because of Theorem 10.5 (iii), A0.27), and A0.28) there is an exact sequence 0 —> H{~1) Here r@) is the projection. 0. A0.34) tt л is the sheaf of relative (n + l)-forms with respect to <p : X -> Д x M. Its support is the critical space С С X of <p. It is a free Oc -module of rank 1. The sheaf QF := A0.35) is a free 0M-module of rank /x. Via the Kodaira-Spencer isomorphism a in A0.3) it is a free 7^-module of rank 1. We write the action as (X, [w]) i-> a(X) ¦ [a] for X e Тм, [со] е fif. A0.36) Vector fields X e TM will be identified with their canonical lifts to Д x M, that is, with the fibrewise global lifts X which satisfy X(z) = 0. Corollary 10.8 For any к € Z one has the following, (i) V : (-k+i'> A0.37)
178 Gaufi-Manin connections for hypersurface singularities (ii) There are exact sequences where r(i+" := r<0) о Vr* O- (Hi) IfX € TM and a e пшН{к) then ^(VxVg-'a) = -a(X) ¦ rlk\a). A0.38) A0.39) (iv) Suppose that v{ e H^ is represented by a volume form u(x, t)dx0 ...dxn € Q"^ with м@) ф 0. Then the period map A0.40) is injective and yields a splitting of the sequence A0.38) for к = -1 Proo/: (ii) follows from Theorem 10.7 (a) and A0.34), (i) and (Hi) follow from the formulas A0.20) and A0.21) for the covariant derivatives of sections in H(~[). (iv) follows from A0.39) and the fact that a volume form represents a generator of f2fi0 as 7^i0-module. ? Remarks 10.9 (a) By definition of H(k\ the pair {Hik\ V) has a logarithmic pole along V, thus besides A0.37) one has V : A0.42) One can also see this algebraically [SK6][SK9], using A0.39) and the fact that precisely the vector fields in DerAxW(logP) are liftable along cp to X (cf. for example [Lo2, F.14)]). (b) A primitive form [SK6][SK9] is a section u, in H^ (or in jr*ft@)) which is represented by a volume form and induces a splitting A0.41) with good properties with respect to K. Saito's higher residue pairings (cf. sections 10.4 and 11.1). Our quite elementary approach to the sheaves H(k) and their properties should be compared with the constructions in [Phi] [Od2] [SM3]. The fol- following theorem comprises the relations. A detailed proof would require precise definitions of all the notions and objects, from which we refrain. We give only a rough sketch. We refer to [Phi] [KK] [Bj] and to the survey articles 10.4 Higher residue pairings 179 [Odl] [No] [Kas2] for the definition of regular holonomic ©-modules, char- characteristic varieties, and good filtrations. We will not subsequently make use of Theorem 10.10. Theorem 10.10 The union {JksZ^k) 's a reSular holonomic VAxM-module with characteristic variety ГдхМ(А x M) U T?(A x M) and good filtration Ti^- Asa ?>дxм-module it is isomorphic to the microlocal Gaufi-Manin system constructed in [Phi] [Od2] [SM3]. Sketch of a proof. The ©-module \Jk€Z H(k) is a local system on A x M - V and singular along V. Therefore its characteristic variety contains the zero section T*xM(A x M) С Г*(А x M) and the conormal bundle 7^(A x M) of V, that is, the closure of T? (A x M). In order to see that this is all, first one has to check that 7i(*' is a good filtration. This follows from Theorem 10.7 (a). Then one has to study the ideal in Ot'm which is defined by that filtration and which determines the characteristic variety. It has homogeneous (in the fibre variables) generators of degree 2 and degree 1. The generators of degree 2 come from the defining relations ^- о ^- = J2k au э7 f°r tne multiplication in TM and from A0.39). The zero set of the ideal of these generators is the zero section and the д + 2 dimensional union of the conormal bundles of all shifts of V along Эг. The generators of degree one come from the logarithmic vector fields. They cut out the zero section and Ц(А x M) (see also [SK9, p. 1254], [Od2, B.3)]). Therefore {JkeZ H(k) is a holonomic Рдхм-module with characteristic va- variety as claimed. It is regular holonomic because the subsheaves Ti1-^ have a logarithmic pole along V. The microlocal GauB-Manin system [Phl][Od2][SM3] is the unique exten- extension of Hm to a VAxM module (J*>o 9*^@) such Ла1 9г is avertible on the germ at 0. Because of Theorem 10.7 (a), the union (J*>o ^(t) *s isomorphic to ? 10.4 Higher residue pairings Let /, F, <p : X -+ A x M, Я", V, and H(k) be as in sections 10.1 and 10.2 with n > 1. In this section we formulate the properties of K. Saito's higher residue pairings on я*7^@). For the algebraic construction and proof we refer to [SK7], for sketches to [SK6] [Nam] [SK9]. Constructions in the framework of microlocal differential systems can be found in [Od2] [SM3]. There it is
180 Gaufl—Manin connections for hypersurface singularities emphasized that Kashiwara observed that the higher residue pairings can be identified with a microlocal Poincare duality. We will offer two completely different descriptions of the higher residue pairings. But both are valid only under some restrictions. The first in Theorem 10.13 holds only for singularities with nondegenerate intersection form, the second in section 10.6, Theorem 10.28, holds only for the restriction to (A x @), 0). Nevertheless both are instructive and explain the properties of the higher residue pairings. For the family of functions Ft, t ? M, and the fixed coordinates xq, ... ,xn on the fibres of the projection ргм '¦ X —*¦ M, (x, t) н-> t, the Grothendieck residue associates to a relative n + 1-form со ? (ргм)*?1'^м a function . Элго' dF Bx. A0.43) It can be defined algebraically or analytically [GrH]. The value at t ? M is 1/B7T(O"+I times the integral of the form <w/ff • • • • • fj- over the cycle {(*, 0 I iff I = К} С B?+l x {t} for a sufficiently "small у > 0. It induces a welldefined pairing Jf:nfxfif-> OM, ... dxn], [g2d*0... dxn]) н» Res дум 3F_ 1? Ъхп A0.44) = (Ргм)*^х/Ахм (cf- (Ю.35)). К. Saito observed that JF is independent of the coordinates xq, ..., xn. By Grothendieck it is nondegenerate. The sheaf QF is not only a free Ом-module of rank fi, but also a free (Prc,M)*Cc-rn°dule of rank 1. By definition, JF satisfies for g € (prc,A/)*Oc, [a>i], [«2] ? ?2F. So, JF is a canonical nondegenerate pairing on QF = 7Г»'И(О)/7Г*'И(~1* (cf. A0.38)). K. Saito extended it to a series of pairings on ntTiP\ his higher residue pairings Кр~к\ Theorem 10.11 ([SK7][SK6][SK9]) There exists an Ом -bilinear pairing KF : n*ni0) x л,П@) -> OM[[Эг-1]] • Э"" A0.46) KF((O{, @2) = / Кр (со\, (О2) ¦ Эг " with Kp~к\ш\, ау?) ? Ом and the following properties. 10.4 Higher residue pairings (i) The pairing Kp~k) is (—1)* -symmetric, (ii) For w\, саг б 181 (Hi) Fora,b < 0, a + b < -k and for u>i,a>2 ? тг (iv) Forш\,шг& KF(z ¦ co\,a>2) — KF(co\, z ¦ un) = [z, KF(co\, 002)], where [z, Эг~*] := к Э"*-1. (v) For ш\,со2 ? nJ-(S~x), X ? TM (X ? TM is lifted to A x M such that X(z) = 0) X KF{co\, сог) = KF{Vxa)\, «2) + KF(tou VX(O2). Remarks 10.12 (a) In view of Theorem 10.7 (a) and the properties of KF, one can extend KF to a pairing on Ujkez^*^'*' ^ш values in ОдЛСЭ^'ШЭг] and similar properties. (b) This extension is studied in [SM3, 2.7]. From the microlocal properties of the GauB-Manin system one obtains KF(m,CO2) ? OMfi{\KX)) ¦ К"'' for«,,u>2 ? nf * (n*H@))Q (c) It is also proved in [SM3, 2.7] that KF is uniquely determined by its properties. This follows essentially from the fact that the germ (/J Ox)o of the GauB-Manin system is simple holonomic as a microdifferential system, so that the only endomorphisms of it are multiplications by scalars ([SM3, 2.7], [Kas2]). The intersection form / on Hn(<p~\z, t), Z) for a regular fibre (p~l(z, t) induces a map Я„(<р~'(г, t), Z) -> Hn(cp-l(z, t), Z), S ^ 1(8,.) [or put the other way, the intersection form is induced by the canonical map Hn(<p~\z, t\ Z) ^ Hn(<p~\z, t), d<p~l(z, t), Z) = H\<p-\z, t), Z) ]. The intersection form / is flat on the homology bundle Я„= (J Hn(<p-\z,t),C)
182 GauB-Manin connections for hypersurface singularities and induces a homomorphism Hn —> H" of flat bundles. If the intersection form is nondegenerate then this map is an isomorphism and induces a flat bilinear form I* on the cohomology bundle H". The residue pairing JF was defined by residues in the fibres of prM : X -»¦ M. We can present the pairings K^k) by residues in the fibres of n : Д x M -> M in the case of a nondegenerate intersection form. The following result is infor- informative, but we will not subsequently use it. The residue at (zo, i) of a meromor- phic function g,(z) in A x {t} is by definition res{Zo,t)(gt) = ?i f\z-zo\=y 8t(z)<k for some small у > 0. Theorem 10.13 Suppose that the intersection form I of the singularity f is nondegenerate. Let I* be the induced bilinear form on the cohomology bundle H". Then for w\, «2 e 7r*H@), fixed t e M, and anyk>0 A0.47) Sketch of two proofs: I can offer two quite indirect proofs, but not a direct proof which would relate A0.47) to K. Saito's definition of KF. The first proof uses a result of Varchenko and the result of M. Saito in Remark 10.12 (c). By [Va6, §3.2, Theorem], the right hand side of A0.47) is equal to JF{[co\], [«2]) for к = 0. One can and has to check that the series of bilinear forms defined by the right hand side of A0.47) satisfies also the other properties in Theorem 10.11. Property (i) holds because I is (— l)"-symmetric. The properties (ii), (iv), and (v) are simple. The first line of (iii) follows from the definition of H^ in Lemma 10.2: for generic t (and thus for all t) the function l*(Vg+kai, co2) is holomorphic on Д x {t} if &>] e Hw, co2 e H{b) with a + b <~-k (cf. [Va6, §3.2]). Because of Remark 10.12 (c), KF must now coincide with the series of bilinear forms defined by the right hand side of A0.47). The second proof will follow from the construction of a Frobenius mani- manifold in section 11.1, from the induced isomorphism between (#", V, I*) and (л*ТМ, V(~ г\ 7<-f' up to a scalar) (cf. chapter 9 for the notations), and from Lemma 9.11. In order to determine the scalar between 7(~t) and /* one again needs [Va6, §3.2, Theorem] D Remarks 10.14 (a) Formula A0.47) explains the properties of the pairing KF, as one could see in the sketch of the first proof. (b) One can obtain an analogous formula for the general case of a singularity with any intersection form if one embeds the Milnor fibration into a fibration with projective fibres as in [Va6, §3.3]. But then one has to check that things are 10.5 Polarized mixed Hodge structures and opposite filtrations 183 welldefined and independent of choices. We refrain from carrying it out here. Theorem 10.13 is informative, but we will not subsequently use it in any case. 10.5 Polarized mixed Hodge structures and opposite nitrations There are several possible definitions of a polarized mixed Hodge structure. The definition presented in [CaK] [He4] is motivated by Schmid's limit mixed Hodge structure. It is the correct one in the case of isolated hypersurface sin- singularities. Steenbrink's mixed Hodge structure [Stn] is polarized in this sense. One can recover it from the Brieskorn lattice. Varchenko [Va2] first saw that the Brieskorn lattice induces mixed Hodge structures. In section 10.6 we will regard this as a feature of the Brieskorn lattice. Here we want to resume the definition of polarized mixed Hodge structures and the construction of classifying spaces for them from [He4, ch. 2]. Opposite filtrations and classifying spaces for them will also be discussed. The choice of an opposite filtration is necessary for the construction of a Frobenius manifold in section 11.1. The weight filtration of a polarized mixed Hodge structure in the sense of Definition 10.16 comes from a nilpotent endomorphism. Its properties are given in the following lemma from [Schm, Lemma 6.4], (cf. also [Gri, pp. 255-256]). Lemma 10.15 Let m e N, Hq a finite dimensional Q-vector space, S a nondegenerate bilinear form on Hq, S : Hq x Hq -> Q, which is (— l)m-sym- l)m-symmetric, and N : #q —>¦ Hq a nilpotent endomorphism with Nm+l = 0, which is an infinitesimal isometry, i.e. S(Na, b) + 5(a, Nb) = Ofor a,b <= Hq. (a) There exists a unique increasing filtration 0 = W_i С Wo С ... С W2m = Hq such that N(Wt) С Щ-г and such that Nl : Gr^+/ -> Gr^_, is an isomorphism. (b) If I +1' < 2m then S(Wh Wr) = 0. (c) A nondegenerate (— \)m+l-symmetric bilinear form Si is welldefined on Gr™+1 for I >0by the requirement: 5/(a, b) = S(a, N% ifa,% e Wm+i represent a,b e Gr^+(. (d) The primitive subspace Pm+i(Ho) of Gr%+1 is defined by Pm+l =ker(Nl+l :Gr^+,- if I > 0 and Pm+, = 0 if I < 0. Then = O^NlPn rm+/ A0.48) A0.49) and this decomposition is orthogonal with respect to Si if I > 0.
184 Gaufi—Manin connections for hypersurface singularities Definition 10.16 [CaK][He4] A. polarized mixed Hodge structure (abbrevi- (abbreviation: PMHS) of weight m is given by the following data: a lattice #z with Нъ С Hq с #r с Не = Hz ® C, a bilinear form S on Hq and an endomor- phismTV of Hq such that от, Hq, S, N, W,, Si, Pm+/ satisfy all properties in Lemma 10.15, and a decreasing Hodge filtration F' on Hq with the properties: (i) The induced filtration F'Gr^ gives a pure Hodge structure of weight к on Grf, i.e. Grf = F"Grf ® F*+|-"Grf, (ii) The endomorphism TV satisfies N(Fp)cFp-\ i.e. N is a (-1,-1)- morphism of mixed Hodge structures, (iii) The Hodge filtration and the pairing satisfy S(FP, Fm+l~p) = 0, (iv) the pure Hodge structure F' Pm+i of weight m +1 on Pm+i is polarized by 5,, i.e. (a) S/(F"Pm+,,F+'+1-"Pm+,)=0, and 03) i2p-'n-lSi(u, Й) > Oif и 6 FpPm+i П Fm*-ppmM, и ф 0. Remarks 10.17 (a) The primitive subspace Pm+i carries a pure Hodge structure, because it is the kernel of the morphism Nl+* : Gr^+; -*¦ Gr^_,_2 of pure Hodge structures. The strictness of the (—1, — l)-morphism N also implies F'Grf = 0 FpNJpk+2j and F'N'Pk+2j = NJFp+* Pk+2J. (b) In Lemma 10.15 the number m could be replaced by some bigger number, but in Definition 10.16 the weight m is essential for condition (iii). Also the assumption that S is (—l)m-symmetric is not important in Lemma 10.15, but essential for (iv) (fi). (c) Condition (iii) implies (iv) (a), but in general it is not equivalent to (iv) (a). One can easily see the following. Under the assumption of all conditions except for (iii) and (iv), condition (iv) (a) for all p and I is equivalent to S(FP П Wm+i, Fm+1-p П Wm-i) = 0 for all p and /. (d) The form S is called a polarizing form. Deligne [De2, Lemma 1.2.8] defined subspaces lp'q for a mixed Hodge structure, which give a simultaneous splitting of the Hodge filtration and the weight filtration. They also behave nicely with respect to a polarizing form ([SM3, Lemma 2.8], [He4, Lemma 2.3]). Lemma 10.18 For a PMHS as in Definition 10.16 let Ip-i :=(Fpn Wp+q) П (J4 П Wp+q + ^ У4'' П VV«-/-i) ¦ A0-5°) 10.5 Polarized mixed Hodge structures and opposite filiations 185 Then q \ Я Fp = G) I1'", I™, A0.51) A0.52) A0.53) р+ч<1 U4~\ (Ю.54) ,r-s) = 0 for(r,s)^(m-p,m-q). A0.55) For p + q >m let Iq'4 be the primitive subspace oflp'4, I+1 : Ipq -*¦ lm~q~x•т~р-1). A0.56) Then S(N4p'q, N4™) = 0 for (r, s,i + j) A0.57) -m). A0.58) Definition 10.19 An opposite filtration for a PMHS as in Definition 10.16 is an increasing filtration U, on He with N(UP) С Up S(UP, Um-i-p) = 0. A0.59) A0.60) A0.61) and I/, = Remarks 10.20 (a) The decomposition A0.59) is equivalent to A0.62) and also to A0.63), see Definition 7.15, pFqnU4, A0.62) i. A0.63) (Ю.64) (b) Condition A0.60) and condition (ii) in Definition 10.16 imply N{FP П Up) С Fp -y Therefore N and all powers NJ of N are strict morphisms for I/., that is, Nj(Up) — ImNj П Up-j. The powers N' are strict morphisms with respect to F* and W, because of Deligne's Ip-q (Lemma 10.18).
186 Gaufi-Manin connections for hypersurface singularities (c) Condition A0.61) and condition (iii) in Definition 10.16 imply S(FpnUp,F4nUq) = 0 for р+дфт. A0.65) The pairing S : F" П Up x Fm~p П Um^p ->¦ С is nondegenerate because S is nondegenerate. (d) If there is a holomorphic family of Hodge nitrations, each giving a PMHS, then an opposite filtration for one PMHS is an opposite filtration for all PMHSs nearby. Being an opposite filtration is an open condition. This follows from codim Fp П Up = codim^P + codim[/p. (e) There exist opposite filtrations. Lemma 10.18 induces a canonical oppo- opposite filtration ?/^°> by /'•«. A0.66) Uq- i<P It will be useful in the discussion of the symmetries of a single singularity in section 13.2. But it will not be useful in the discussion of д-constant fami- families of singularities in chapter 12. In general the Hodge filtration varies there holomorphically. One can check that the opposite filtration t/<0) varies antiholo- morphically. But we need a constant opposite filtration. Lemma 10.21 The space of all opposite filtrations as in Definition 10.19 for a PMHS as in Definition 10.16 is an algebraic manifold isomorphic to CN°" for some Nopp € N. Proof. The algebraic group [g € Aut(#c, S, N, F')\g = id on Gr? for all p\ A0.67) is unipotent. We will see that it acts transitively on the space of all opposite filtrations. Then this space is isomorphic to №' for some Nopp € N (cf. for example [Bore, 11.13]). One starts with a basis af\ ..., a™ of Hc (here ц. := dim Hc) which fits to Deligne's Ipq, that means, a^ e I^W) for each / and some p(i), q(i) e Z. It fits to the decomposition Hc = 0p Fp П Up°\ that is, a\0) € Fpii) П Up^iy O h One has Nar = and A0.68) for some v,7, au € C. If U. is any opposite filtration, then there are canonical isomorphisms Fp n A0.69) 10.5 Polarized mixed Hodge structures and opposite filtrations 187 There exists a unique basis a\,..., aM of He which satisfies щ 6 Fp(i) П Upd) and A0.70) We claim that ai,..., aM satisfy precisely the same relations with respect to N Naj = 2^ VijOj and S(at,aj) = air j A0.72) The relation for N follows from p(j) = p(i) — 1 for v-tj / 0, from iVa,- - ? v,va; 6 t/p@-, = iV (a,- - a?») - ^ vy (a; - af) € F"«, j J A0.73) and from Fp@ П f/p(l-)-i = {0}. The relation for S follows from A0.61) and condition (iii) in Definition 10.16. The automorphism of #c which maps a] to a,- is an element of the group in A0.67) and maps t/^0) to U,. Therefore the group in A0.67) acts transitively on the space of all opposite filtrations. D Often the situation is more complicated. In the singularity case one has the following setting: (i) a lattice Hz С HQ С HR С H - Hc = Hz ®z C; (ii) a quasiunipotent monodromy h = hs ¦ hu on Hi with semisimple part hs, unipotent part hu, and nilpotent part N = loghu on tfQ; the eigenspaces H\ = кег(й5 — к : H -*¦ H); a weight filtration W. on Hq whose restrictions to Hi and Нф\ := ©x#i ^ come from N as in Lemma 10.15 and are centred at n + 1 and n; (iii) a nondegenerate bilinear form S on Hq (respectively on Hz with values in Q) whose restrictions to Hi and Нфх are (-1)"+1- and (-1)"-symmetric; S is monodromy invariant; then N is an infinitesimal isometry; (iv) an /^-invariant Hodge filtration Fo* whose restrictions to Hi and H^\ form PMHSs of weight n + 1 and n together with the other data. We also call this sum of PMHSs on Hi and H^i a PMHS. Deligne's Ip>4 and the I™ are hs-invariant and decompose into eigenspaces {lp-q)x and (/,f'% of hs. One should consider only monodromy invariant opposite filtrations. Condi- Condition A0.61) has to be taken with m = n + 1 on H\ and m — n on H^i. Lemma 10.21 also holds for such opposite filtrations.
188 Gaufi-Manin connections for hypersurface singularities In [He4] the space Dpmhs := {filtrations F' on Hc \ dim. FpPlx = dimFopP;iX, F* is /jj-invariant and induces PMHSsof weight и + 1 and и on Hi and Я#1} A0.74) is defined. It is a classifying space for Hodge filtrations F* giving PMHSs with the same dimensions as Fq . It is studied together with the group Gz := Aut(#z, h, S) and other spaces and groups. A0.75) Theorem 10.22 [He4] In the above setting, the classifying space Dpmhs is a real homogeneous space and a complex manifold. The group Gz acts properly discontinuously on it. The moduli space Dpmhs I'Gzfor isomorphism classes of polarized mixed Hodge structures as above is a normal complex space and has only quotient singularities. Remarks 10.23 In [He4] a bigger space DPMhs is also defined. It is the space of the filtrations F* which are similar to Fo* with respect to all conditions except those involving complex conjugation. Dpmhs is an algebraic manifold and a complex homogeneous space. The group Gq — Aut(#c> h, S) acts transitively on it. It is a bundle over a product Dpnm of projective algebraic manifolds with fibres isomorphic to C"^" for some number Njjhre e N. The space Dpmhs is the restriction of this bundle to an open submanifold Dprim С Dprim, which is a product of classifying spaces for polarized pure Hodge structures. 10.6 Brieskorn lattice We use the notations from sections 10.1-10.3; the function f(x0,...,, xn) is an isolated hypersurface singularity with и > 1. Its Brieskorn lattice Щ is the germ at 0 of a distinguished extension to 0 of the sheaf of holomorphic sections of the cohomology bundle H" | д. x (o) of a Milnor fibration for /. It was defined as the set of germs of sections of Gelfand-Leray forms of (и + l)-forms, Щ = ttgl, 0/d/ л йп"с~1 0 (cf. A0.18) and A0.25)). Its relation to the GauB- Manin connection fora semiuniversal unfolding and some of its properties were already formulated in Theorem 10.5. The Brieskorn lattice is a key to the singularity. It has a rich structure and a long history, which started with [Bri2]. Here we give a concise report on its structure, but not on its history. We follow [He4]. 10.6 Brieskorn lattice 189 The notions of sections 7.1 and 7.2 (e.g. the spaces Ca of elementary sections and the V-filtration) will be used freely. The space H°° is the At-dimensional space of multivalued global flat sections on the cohomology bundle Я"|д.х|0) over the punctured disc Д* (cf. G.2) and G.3)). Now it contains the lattice #°° = Z" of multivalued sections of cohomology classes in Я"(/~'(г), Z). Here f~\z) is a Milnor fibre of the representative / : В^+|П/-'(А)-^ Д of/. In addition to the monodromy h on H°°, there is a nondegenerate bilinear form S : H™ x Я|° -» Q [He4]. It is less well known than it deserves. It makes Steenbrink's mixed Hodge structure [Stn] on H°° into a polarized mixed Hodge structure. It induces a series of bilinear forms on НЦ which coincide with the restriction to #o °f K- Saito's higher residue pairings. It is defined in A0.78). The variation operator Var : Hn(f~l(z),Z) -> Hn(f~\z), Z) for a Milnor fibre /"'(г) uses the identification Я"(/~'(г), Z) = Я„(/"'(г), 9/-'(г), Z) and maps a relative cycle у to the absolute cycle h(y) — y, where h is a geometric monodromy on the Milnor fibre which fixes the boundary of it. Var is an isomorphism and determines the monodromy h and the intersection form / on Яп(/-'(г), Z) (cf. for example [Lo2][AGV2]). The monodromy has a semisimple part hs and a unipotent part hu with N :— log hu. It acts on Hn (/-' (z), Z), Я"(/"' (z), Z), and H§°. We use the notations Я~ = кег(/г, - A.) and H? = ф^, Ял°°, and the same for ЯЛ(/"'(г), С) andЯn(/-1(z),<C). The intersection form / on Я„(/~' (г), C)^i is nondegenerate. It induces an isomorphism to Я"(/"' (z), C)^i and a form /* on it. An isomorphism v : Яп(/-'(г), Q) -*¦ Я"(/"'(г), Q) is defined by опЯп(/-1(г),<С)_,ь (Ю.76) on The polarizing form 5 on Я"(/~Чг), Q) is defined by S(a, b) = (-l)n("-1)/2(a, Var о v(b)). A0.77) A0.78) Lemma 10.24 [He4] The form S is nondegenerate and monodromy invari- invariant. The restriction to Ял(/"'(г), C)^i is (— 1)" -symmetric and is equal to (_l)«(«-i)/2/* The restriction to Я"(/"'(г), C)i is (-l)"+l-symmetric. Remarks 10.25 Unfortunately, in [He4] it was not noticed that the monodromy there is the inverse of that in [Schm]. Therefore here v and 5 on Я" (/"' (z), C)i
> г' ¦) >"¦'•'ii 190 Gaufi-Manin connections for hypersurface singularities differ by a sign from those in [He4]. Also, in A0.83) a sign turns up, and in Theorem 10.30 (i) one needs —N. The form S on Hn{f~\z), Q) induces a form S with the same properties on #0°. This form S can be used to define a C{{9.T1}}-sesquilinear pairing Kf : 3-1 A0.79) on the space V»-' = ©_1<ff<0C{z}Ca = ф_|<а<0С{{Эг-'}}Са (cf. section 7.1). One needs the homomorphisms ф-а in G.6) from multivalued flat sections to elementary sections, Са, es(A, a), A0.80) in order to go from H°° to V >-1. The pairing Kf and its parts Kf~k) with values in С are defined by the formulas A0.81)-A0.85), here a, p e (-1,0], a € Ca, Kf{a'b) = •9 -l for a = p = 0, a, b), (fk\ Kfigia, g2b) = J^ K(fk\g\a, g2b) A0.81) A0.82) A0.83) A0.84) A0.85) Lemma 10.26 [He4] (i) The pairing K(f k) is (-1)*-symmetric. (ii) For the restrictions ofKf to the eigenspaces one has Kf : С хС' —>¦ 0 fora+p <?Z,a,p> -1, and A> :CaxC^ С-ЗГ"^ is nondegenerate fora+peZ,a,p > -1. finj Г/ге pairing Kf and the multiplication by z are related by Kf{z a, b) — Kf(a,zb) = [z, Kf{a, b)\ where [z, д~к] = кд~к-\ The pairing Kf shares the properties (i), (iii), and A0.84) with K. Saito's higher residue pairings (Theorem 10.11). Theorem 10.28 will show that it co- coincides on Щ with the restriction of KF to Я„. 10.6 Brieskorn lattice 191 Now we come to the properties of Щ (cf. A0.18), Theorem 10.5, A0.25)). Brieskorn [Bri2] showed that its germs of sections have moderate growth (that means Щ С V>~°°), that they generate the cohomologies of the Milnor fibres and that the map V8; : Щ -+ Щ from the sublattice Щ С Щ defined by n-forms is an isomorphism (cf. A0.17), Theorem 10.5 (v)). Malgrange [Mall] showed Hq с V>0. Together with Lemma 7.4 this gives the following. Theorem 10.27 (i) The Brieskorn lattice Щ is a free C[z}-module of rank \x, equivalents C{z}[z']]H^ = V^00. (ii) The Brieskorn lattice and Щ satisfy Щ С V>0, H? С V>-\ Щ = ^Щ С Щ. (Hi) The Brieskorn lattice Щ is a free C{{3~' }}-module of rank ц. The nondegenerate Grothendieck residue pairing Jp : ?2f x Qf —> Ом from A0.44) for the semiuniversal unfolding F of f induces a nondegenerate pairing Jf : on the /x-dimensional space A0.86) A0.87) The following theorem from [He4] gives the relation between Hg, Kf, J/, and KF. Remarks on the proof will be given afterwards. Theorem 10.28 (i) For к > 0, the pairing Kf~k) on Hg is the restriction of К Saito's higher residue pairing К^к) on Tif (cf Theorem 10.11) to Hg. (ii) The pairing and the Brieskorn lattice satisfy Kf~k\Hg, Hg) = 0 for -n<k< -1, i.e. Kf(Hg, Hg) = CU3-1}} • Э-"-1. (iii) The pairings Kf* and Jf are related by Kf\coi,aJ) = J/dm], {(&Л) for a>\, o>i € Hg. (iv) The Brieskorn lattice Hg is isotropic of maximal size with respect to the antisymmetric bilinear form K^l\ that means, a section a> € V>-1 satisfies Kfl\co, Hg) = 0 if and only if со е Hg. (v) Therefore Hg D Vn~l and dim Hg/ V"~l = |dim V^/V'K and the spectral numbers ai,..., aM are contained in (— 1, n) and satisfy the symmetry _,- = n — Remarks 10.29 (a) Part (i) can be found essentially in [SM2, Appendix], [SM3, 2.7]. But he is not very explicit about the definition of Kf. Another
192 Gaufi-Manin connections for hypersurface singularities way to check it is with the results in [Va6, §3.3]. Varchenko embeds the Milnor fibration (and its extension to the semiuniversal unfolding) into a fibration with projective fibres. He relates the Grothendieck pairing J/ (and Jf) to the prim- primitive cohomology bundle of this extended fibration and to the nondegenerate intersection form on it. He considers formulas of the same type as A0.47), but for the primitive cohomology bundle. I checked that these formulas induce a bilinear form on Ti0 with the same properties as K?, that they induce Kf on V>~1, and that the properties (ii)-(v) hold. Now one needs M. Saito's result ([SM3,2.7], see Remark 10.12 (c)) that the form Kf on 7Yq0) is uniquely determined by its properties. One obtains (i) and a generalization of Theorem 10.13 (see Remark 10.14 (b)), (b) The existence of a relative compactification of the Milnor fibration to a fi- fibration with projective fibres and other good properties is due to [Bri2][Schel]. It is not only essential for the proof of Theorem 10.28, but also for that of Theorem 10.30. (c) Part (iv) follows from the facts that // is nondegenerate and that Kf is CUa-'H-sesquilinear (cf. A0.84)). The form Kf satisfies Kf[)(V>-\ Vй) = 0 and induces a nondegenerate form on the quotient V>~x/V". This shows the inclusions V С Щ and ai,..., ад e (—1, n) and the di- dimension formula in (v). A refinement of this argument yields the symmetry of the spectral numbers. Varchenko [Va2] [Val] [AGV2] showed that the principal parts of the sections in Hq induce a holomorphic family of Hodge filtrations on the cohomology bundle, giving a mixed Hodge structure on each fibre. There is a limit Hodge filtration FVa on H°°. He used a relative compactification of the Milnor fibration and results of Griffiths, Schmid, and Scherk. His construction was modified [SchSt] [SMI] [Ph3] to obtain Steenbrink's [Stn] mixed Hodge structure F* on Я°°. It is polarized by S because of an exact sequence of Steenbrink, connecting it with the limit mixed Hodge structure of Schmid for a relative compactification of the Milnor fibration. This was emphasized in [He4] (for the sign in — N in part (i) see Remark 10.25). Theorem 10.30 (i) The subspaces FPH°° := *"' G.27)) A0.88) for a e (—1,0], e~lma — X, define an hs-invariant decreasing filtration on H°° with 0 = Fn+i С F" С ... С F° = H°°. It is Steenbrink's Hodge filtration. 10.6 Brieskorn lattice 193 Together with S and -N it gives a PMHS of weight n on Hg\ and a PMHS of weight n + 1 on Hf° (in the sense of Definition 10.16). (ii) The subspaces HaH? ¦= ir;xz-n+'Grav+n~PHZ (= Fn-Pfwm G.24)) A0.89) define an hs-invariant decreasing filtration on H°°. Together with the weight filtration W. from N this filtration also gives a mixed Hodge structure. The filtrations Fya and F' coincide on the quotients Gr,w (cf. Lemma 7.4 (b)). Remark 10.31 In [AGV2][Stn] /j. spectral pairs in Q x Z are defined. The multiplicity of (a, /) e Q x Z as spectral pair is d(a, I) := dimGr[;-a]Gr?[a+1]tf; oo X A0.90) for e~2"'a = X. Because of the PMHS and the strict morphism N, they satisfy the symmetries (any two of the symmetries determine the third) d(a, l) = d(n-l- a, 2n - I), d(a, I) = d{2n -I-I-a, I), A0.91) A0.92) A0.93) The first entries of the spectral pairs are the spectral numbers. It is unlikely that one can recover all the symmetries only with Theorem 10.28, the strictness of the morphism N is probably too profound. The following result of Varchenko will be used in chapter 12. Theorem 1032 [Va3] The spectral pairs are constant within a ^-constant family of singularities. The next result of M. Saito is essential for the construction of a (good) primitive form in section 11.1. The д-dimensional space Q/ from A0.87) is a free module of rank 1 over the Jacobi algebra O/J/. Generators for this module are represented by volume forms in fi?l"+i 0. that is, forms u(x)dxo ... dxn with м@) ф 0. The space Я/ inherits a finite decreasing filtration V'Q/ from the In- Infiltration on Hq. The dimension dimGr^?2/ = d(a) is the multiplicity of a as spectral number. Parts (ii) and (iii) in Theorem 10.33 are corollaries of part (i). Theorem 10.33 [SM4, 3.11 Remark] (i) The maximal ideal m/J/ in the Jacobi algebra maps VaQ.f to V>aQ/. (ii) «i < a2, that means, d(a{) = 1 and dim Gr^""'1^0?,,,., = 1.
194 GauB-Manin connections for hypersurface singularities (Hi) A form u(x)dx0 ...dxne Q"^}, 0 represents a section in Щ with principal part in C"] if and only if со is a volume form, that is, и@) ф О. In [He4, ch. 5] a classifying space for Brieskorn lattices is studied. Its ele- elements are subspaces ?0 С V> with the following properties. (or) The subspace jCq is a free C{z}-module of rank fi. 03) The subspace ?0 is a free C{{3-'}}-module of rank /л. (у) The decreasing /^-invariant filtration F?J. on H°° from G.24) (cf. A0.88)) is in the classifying space DPMHS for PMHSs from A0.74). (<S) One has K(f~k\?.o, ?o) = 0 for ~n < к < -1. Theorem 10.34 [He4] The classifying space Dbl = {А С V*-1 | ?0 satisfies (a), 03), (y), (S)} A0.94) for Brieskorn lattices is a complex manifold and a locally trivial holomorphic bundle prBL : DBL -*¦ DPMhs with fibres isomorphic to CN"L for some NBL 6 N. The fibres have a natural C*-action with negative weights. The group Gz = Aut(#?°, h, S) acts properly discontinuously on DPMHS and thus also on DBL. More details on DBL including a formula for NBL can be found in [He4]. Essentially because of Theorem 10.32, there is a period map from the ^-constant stratum to the space DBL. Such period maps are studied in [Hel] [He2] [He3] and in section 12.2. Chapter 11 Frobenius manifolds for hypersurface singularities The construction of Frobenius manifolds for singularities is due to K. Saito and M. Saito, using results of Malgrange. The version presented in section 11.1 replaces the use of Malgrange's results by the solution of the Riemann-Hilbert- Birkhoff problem in section 7.4 and by the tools in section 8.2. All the other ingredients from the GauB-Manin connection are provided in chapter 10. Section 11.2 establishes series of functions which are close to Dubrovin's deformed flat coordinates. Some use of them is made in chapter 12. In view of some results of Dubrovin, Zhang, and Givental one can hope that much more can be found in these series of functions. Sabbah generalized most of K. Saito and M. Saito's construction to the case of tame functions with isolated singularities on affine manifolds [Sab3][Sab2] [Sab4]. But the details are quite different, there one uses oscillating integrals, and the results are not as complete as in the local case. We discuss this at some length in section 11.4. The case of tame functions is important for the following question within mirror symmetry: Are certain Frobenius manifolds from quantum cohomology isomorphic to certain Frobenius manifolds somehow coming from functions with isolated singularities? This is motivated by the results of Givental. A special case was looked at by Barannikov. In section 11.3 we make some remarks about this version of mirror symmetry. 11.1 Construction of Frobenius manifolds Let / : (C+1, 0) —> (C, 0) be a holomorphic function germ with an isolated singularity at 0, with Milnor number ц,, and with n > 1. The base space of a semiuniversal unfolding is a germ (M, 0) = (<СД, 0). It can be equipped with the structure of a massive Frobenius manifold. The multiplication and the Euler field are unique, but the flat metric depends on some choice. Its existence 195
196 Frobenius manifolds for hypersurface singularities is highly nontrivial and follows from the existence of a primitive form of K. Saito [SK6][SK9], which was proved in the general case by M. Saito [SM3], building on work of K. Saito and many other people. M. Saito used a result of Malgrange [Mal3][Mal5] on the extension of special bases in the microlocal GauB-Manin system, whose proof involved the Fourier-Laplace transforma- transformation of this system. That made it difficult to use the construction of the metrics on (M, 0) and to work with these Frobenius manifolds. Below we give a simpler and more explicit version of the construction, with- without using Malgrange's result. We will also provide precise information as to which choices have to be made and what they yield. This makes the construction sufficiently transparent to be subsequently applied. To describe the choices, we need the space Я°° of multivalued global flat sections on the cohomology bundle of a Milnor fibration. It is defined as follows (cf. G.2) and G.3) and section 10.6). Choose e>0 and S >0 as in section 10.1 such that f:Bne+x Л /-'(A) ^ A for A := B\ С С and 5e"+1 С C"+1 is a Milnor fibration. The cohomology bundle Я"|д. := ЦбД. Hn{f~\z), С) has rank ц and a flat structure. The universal covering e : С —> С*, f \-> e2*'* restricts to a universal covering e : e~l(A*) -*¦ A*. If pr : е*(Я"|д.) -»• Я"|д. denotes for a moment the projection, then H°° = {proA\ Aisa global flat section of e*(Hn\A.)} A1.1) is the /^.-dimensional space of multivalued global flat sections of the cohomology bundle. It is independent of the choice of the Milnor fibration. It is equipped with a lattice Я^° С Я°°, a monodromy A with semisimple part A, and unipotent part А„, a polarizing form S (see section 10.6), and a polarized mixed Hodge structure with Steenbrink's Hodge filtration F* (see Theorem 10.30). The spectral numbers au ..., aM e (—1, n) П Q come from the Brieskorn lattice Hq, but they also encode the dimensions of the spaces GrpFH?° (see A0.50)), where Ях°° = ker(A, - X). M. Saito's result (Theorem 10.33) that the smallest spectral number <*i has multiplicity 1 implies 'a"ff% = 1. (П.2) The notion of an opposite filtration for F* was defined in Definition 10.19. Monodromy invariant opposite filtrations U. for F* exist. They have to satisfy A0.61) with m = n + 1 on Я,00 and with m — n on Я?} : They form an algebraic manifold isomorphic to C^"» for some Nop (Lemma 10.21). Such a filtration satisfies )°° fJOO dim = 1. (П.З) 11.1 Construction of Frobenius manifolds 197 Theorem 11.1 Any choice (t/., y\) of a monodromy invariant opposite fil- filtration U, for F' on H°° and of a generator y\ of the 1-dimensional space GrP,_a ]Я°?2».-«, induces the structure of a germ of a Frobenius manifold on (M, 0). Multiplication o, unit field e, and Euler field E are unique. Different choices of (I/., y\) Sive different metrics g. If (?/., y\) gives a metric g then (?/., cyOfor с е С* gives the metric eg. Let Vg be the Levi-Civita connection of g. The flat endomorphism VgE : Тм о -* Тм.О* X i-»- V^jE, is semisimple with eigenvalues dt = 1 + a\ — a,-, i = 1 (i, and D = 2 - (aH - c*i) = 2 + 2a, - (n - 1). This follows from the results of M. Saito [SM3] and K. Saito [SK6][SK9] on the GauB-Manin connection and the primitive forms. In this section we give a simplified proof. It uses ingredients from all previous chapters. A rough idea of the proof was given in section 6.1. Proof of Theorem 11.1: The Brieskorn lattice H? (cf. A0.18), Theorem 10.5, A0.25), section 10.6) induces Steenbrink's Hodge filtration F* = F°ig. on Я00 (Theorem 10.30). Here F^l, is the increasing filtration which is associated to Щ by G.27). Let ([/., y\) be as in Theorem 11.1. Define a monodromy invariant increasing filtration U. on Hc by Up := Up+n on Я,°° and Up := Up+n+\ on Я~. Then F?s and U. are opposite in Я,00, and F?8 and t/.+i are opposite in Н°?{, in the sense of Definition 7.15. The cohomology bundle Я" | д. extends uniquely to a flat bundle over C*. The Brieskorn lattice Hq gives an extension of its sheaf of sections over 0. The filtra- filtration U. induces an extension over oo. By Theorem 7.17 this twofold extension Я" is a free Or -module of rank ц. with a logarithmic pole at oo. The residue en- endomorphism at oo has eigenvalues -a i,..., aM (Theorem 7.17) and is semisim- semisimple because of N(UP) С Up-\ and Theorem 7.10 (N = loghu as usual). Let F : X ->• A be a semiuniversal unfolding of / as in section 10.2 with discriminant Ъ С A x M and cohomology bundle Я" over Ax M -t>. The cohomology bundle Я" extends uniquely to a flat bundle over С x M - V. The twofold extension ~H" embeds into a twofold extension Я@) of the sheaf of sections of Я": The sheaf Ti{0) from Lemma 10.2 is the unique locally free (Theorem 10.5)-extension over V with a logarithmic pole along Vreg with (semisimpleforn > l)residueendomorphismwitheigenvalues(i^!-, 0,..., 0). By Theorem 8.7 U. or (fljf)oo induce a unique extension over {oo} x M with a logarithmic pole along {oo} x M. A priori № is a locally free Or хм-module of rank д. But the restrictions of its sections to P1 x {0} give the free OPi-module Я". By a classical theorem
1TW . •¦"! 198 Frobenius manifolds for hypersurface singularities on families of vector bundles over P1 (e.g. [Sab4,15.b], [MaI4, §4]) then ТЩ itself is a free OpixM-module if M is small enough. Choosing M arbitrarily small does no harm. So we will always suppose that it is small enough and often go from the germ (M, 0) to such a representative M. As in chapter 9, pr : P1 x M -*¦ M denotes the projection. The sheaf pr^H®) of fibrewise global sections is a free C^-module of rank /x. The restriction to {oo} x M yields a canonical isomorphism of O^-modules Pi (oo)xAf A1.4) The latter sheaf is equipped with the flat residual connection with respect to the coordinate | (see section 8.2). The global sections whose restrictions to {oo} x Mare flat with respect to this residual connection form a/U,-dimensional space. In order to study the features of these special sections we need some choices and notations. The opposite filtration U. induces a splitting of the Hodge filtration F' on H°° with specific properties (cf. A0.59), A0.64), A0.65)). This can be shifted to the spaces C" of elementary sections in Я" | д.х(о) with the maps т/га : Н^„ -> С" from G.5). The subspaces Ga := П ^«) С Са A1.5) have dimensions d(ce) = multiplicity of a as a spectral number. They satisfy (Kf is the pairing on V>~1 from section 10.6): (zV9z -a)Ga = ~Ga С VdzGa+x с A1.6) p>0 Kf(Ga,Gli) = 0 fora+/8#n-l, A1.7) A1.8) °Ga+P = ($z-pG°+i> fora>-l, A1.9) -PGa+p. A1.10) p<0 p<0 In order to describe the situation along {oo} x M we make the following observations: (I) The space H°° is canonically isomorphic to the space of multivalued flat sections on the restriction to (C — Д) x M of the (extended) cohomology bundle Hn. 11.1 Construction of Frobenius manifolds 199 (II) On Hn\(C — Д) x M one has elementary sections with respect to the coordinate z. They are defined by the same formula, compare G.5) and (8.13). An elementary section a of order a satisfies (zV3l-a)"+1or =0 and Vxct=0, A1.11) where X e TM is lifted to P1 x M in the canonical way with X(z) = 0. (III) The elementary sections over Д * x {0} and those in (II) glue to elementary sections over C* x {0} U (С-Д)хМ. From now on the spaces С denote spaces of such global elementary sections. But we will still take the liberty to embed the С into spaces of germs of sections at (Д x {0}, 0) or (P1 x M, (oo, 0)). One will usually see from the coefficients what is meant. By the construction of the extension over {oo} x M, one has - Д) x M = A1.12) Therestrictionto{oo} x M yields a canonical isomorphism of free Cj^-modules of rank fi \ a \ I Z /l(oo}xM. By definition of the residual connection on the right module, the flat sec- sections of this residual connection correspond to the sections in фа Ga on the left. Now we choose a basis s\,..., s^ of фо Ga with Si e G°", d-"-1, A1.14) A1.15) Я™*,,,,. A1.16) Note that dimG = 1 and that s\ is uniquely determined by A1.16). The sections v, := A1.17) for i = 1,..., [i form a basis of the space of global sections whose restrictions to {oo} x M are flat with respect to the residual connection at {oo} x M. Their germs at (oo, 0) satisfy A1.18) here ? := j with "z^j = -zdz.
200 Frobenius manifolds for hypersurface singularities In the following, if X e TM we will also denote by X the canonical lift to P1 x M with X(z) - 0; this lift also satisfies [X, dz] = 0. The operator V^' is welldefined for sections in С with а ф — 1 and, by Theorem 10.7, for sections in H^ and ntHi0\ where ж : A x M -> M is the projection. By the proof of Theorem 10.7, V^1 is also welldefined for fibrewise global sections T^ A1.19) A1.20) Lemma 11.2 (a) For X eTM (b) z ¦ vt € ф OM ¦ vj + (a,- + ' vt. Proof, (a) By Corollary 10.8,7Г*7^@) and the sheaf of fibrewise global sections of ?t@)|cxM are invariant under V^V^. For the germs at (oo, 0) one has to observe the following. The germ 7^0)(oo,o) has a logarithmic pole along {00} x M. It contains C°, because its spectral numbers are —ct\ —aM € (—n, 1). If the monodromy has eigenvalue 1, i.e. С0 Ф 0, then the germs in z" • W@)(oo,o) do not have unique preimages under V3j, but any preimage is contained in W^oo.o)- Therefore the germs at @0,0) satisfy '-I. A1-21) A1.22) Thus Vy V^"' «,¦ is a global section in W®. (b) One has to apply zVdz - (сц + 1) id to A1.21) and observe A1.6). Then at @0, 0) z ¦ Vi - (a,- + ' Vi € A1.23) This section is in any case in 7^@)|cxM- Therefore it is a global section in Wl D In M. Saito's approach, the equations A1.19) and A1.20) follow directly from Malgrange's result. Formula A1.19) is used to prove Lemma 11.1 (b) (cf. [SM3, 4.3]). Again KF is K. Saito's higher residue pairing (Theorem 10.11), and Kf is the pairing on V>-i defined in section 10.6. 11.1 Construction of Frobenius manifolds 201 Lemma 11.3 (a) The restrictions vf € Щ of the sections vt satisfy A1.24) >a,- p>0 9-n-i_ (b) A1.25) A1.26) Proof, (a) Formula A1.24) is obtained as G.44) in the proof of Theorem 7.16: because of A1.9) and A1.10) Щ intersects the right hand side of A1.24) in a unique element. This element extends to a global section in H" and must coincide with the restriction of v-,. The formulas A1.7), A1.8), and A1.24) show Kf{vl v»+w) € SU ¦ Э-" + ?C ¦ Э *i -* A1.27) On the other hand, КГ(Щ, Щ) = CU3,-1}} • Э^" (Theorem 10.7). (b) One considers K^k\vi, vM+i_j) € Ом,о as apower series. The constant term is K{fk)(v^ i^+1_,.) = hj ¦ ho because of Theorem 10.28 (i). The follow- following calculation shows that the higher terms vanish. The third step uses A1.19) once. Here X\,...,Xm € TM$ and k > 0. (X,... Xm K(fk\vi, Vj))@) A1.28) = Kf m-l e /=0 D Remark 11.4 In the case of a singularity with nondegenerate intersection form the description of Kp in Theorem 10.13 gives a better idea why A1.26) holds: Kp~k\vj, Vj) is by A0.47) the residue along {00} x M of a certain meromorphic 1-form on P1 x M. One can check that this residue is constant because of the properties of v-, and Vj along {00} x M.
202 Frobenius manifolds for hypersurface singularities In the case of a degenerate intersection form one can embed the generalized Milnor fibration of the semiuniversal unfolding in a relative compactification and obtain an analogon of Theorem 10.13 (cf. Remarks 10.14 (b) and 10.29 (a)). I checked that one can make a refinement of the choice (?/., y\) and that one can also explain A1.26) in this case with a residue along {oo} x M, but carrying out the details is quite intricate. The section v\ plays a special role. It turns out to be a primitive form in the sense of K. Saito. The germ v° 6 Щ has the principal part in C1. Then by Theorem 10.33 the germ vi 6 П^ is represented by a volume form u(x, t) dx0 ... dxn with м@) ф 0. By Corollary 10.8 (iv) it induces an injective period map v.% M,0 x м- -v ¦V3>b such that Because of A1.19), v is then an isomorphism Let A1.29) A1.30) A1.31) Si:=v-\Vi). A1.32) By the Kodaira-Spencermap a : TMfi -* Oc,o from A0.3) we have the multi- multiplication о on TM0 and the Euler field E = a~'([^])- Lemma 11.5 concludes the proof of Theorem 11.1, except for the statement that different choices (?/., y\) lead to different metrics g. This will be proved after Corollary 11.6. Lemma 11.5 (a) The 5,- satisfy S{ = e and [Si, Sj] = 0 and A1.33) -> SlFi0fromA0.38) A1.34) The Gmthendieck residue pairing JF on ?lF induces a flat metric g on M. The vector fields St are flat with g(Si, <5M+w) = Sy. Let V* denote the Levi-Civita connection of g. (b) The composition ofv and of the projection r@).: coincides with the isomorphism Тм,о-+ nF,0 X ^ a(X) ¦ r 11.1 Construction of Frobenius manifolds 203 (c) The metric g is multiplication invariant and satisfies together with the multiplication the potentiality condition. (d) The Euler field E satisfies [S(, E] = A + «i - а,->5;. Proof, (a) Crucial for (a) and (b) is formula A0.39), r@)( „VxV^'uj) = a(X) ¦ r@)(w) A1.35) for X 6 TM,0, со 6 Wf,0). For со = Vi and X = ^ it shows aE,) = 1, so 5] = e. Together with a(S,- о S,-) = aE,) • aEy-) it shows that A1.33) holds modulo kerr@) = V^0' = П(о~1). Then A1.33) holds because of A1.19) and A1.30). The vector fields 5,- commute because of 7-2. = V(S; О Sj) - v(Sj oSi) = 0. A1.36) (b) The first statement is A1.35) for со = v\. The pairing K^ induces JF via the projection r@) (Theorem 10.11). Then A1.26) shows g(Sh 5M+,_y) = Su. The metric g is flat because the vector fields 6, commute. (c) The metric g is multiplication invariant because JF satisfies the corre- corresponding property A0.45). The manifold (M, o, e) is an F-manifold and e is flat. In Theorem 2.15 it is proved that these three properties imply potentiality. But the potentiality in the form V|Ey о Sk) = Vf.(<5; о Sk) also follows from rewriting 0 = VKiJjlV^2u(l5Jt). Here one needs the OM,0-linear extension of A1.33), -VxV3-1l;(y)=:v(Xoy)-V3;1i;(V|r) A1.37) for X, Y 6 TMfi. (d) The vector fields E - ze and Эг + e and therefore also E + zdz are loga- logarithmic along V by A0.9); in fact, E + zdz is also logarithmic along {oo} x M. The sheaf Vg"'W@) = H(~l) has a logarithmic pole along V. Therefore V?)V^1 Vi € V^Hi0)= z-Vi-v(Eo Si). A1.38) Then A1.20) shows ~ = v(E о Si) + (a; + 1)V8~V A1.39) We write E = ?y SjSj for some ey e OMfi. Comparison of A1.39) and shows J J Si{e}) - 0 for i ф j and 5j(e,) = 1 + ai - a,-.
204 Fmbenius manifolds for hypersurface singularities Therefore V| E = [Sh E] = A + a, - a,)<5;. One also now obtains the following nice formula as an Ом-linear extension of A1.39). For X eTM Z ¦ v(X) = v(E oX)+ + a,)X - A1.41) a Corollary 11.6 The period map v yields an isomorphism between the pairs Gi@>, V) and (рг*Тм, V(~ ^) (the second one is a second structure connection of the Frobenius manifold (M, о, е, Е, g)). Proof. One compares (9.10) and (9.11) with A1.37) and (9.13) with A1.41). One needs V - \ - § = V*E - f - | - § = V«? - B + at). D Nowwecan prove the injectivity of themap {choices (U., yi)} -» {metrics g] in Theorem 11.1. If one fixes U. then the construction of ы in A1.16) and A1.17) shows that the metric g varies linearly with the element y\- Suppose that two choices of (?/., y\) lead to the same metric. By Corollary 11.6, there exists an automorphism of the flat cohomology bundle which maps one extension 7^°> of its sheaf of sections to the other. The auto- automorphism commutes with the whole monodromy group. The monodromy group is generated by the Picard-Lef schetz transformations of a distinguished basis of vanishing cycles. The automorphism commutes with them and must map each distinguished cycle to a multiple of itself. The Coxeter- Dynkin diagram is connected [La] [Ga]. Therefore the multiples are all the same, and the automorphism is a multiple of the identity. Thus the two extensions H<°) coincide and the two opposite filtrations coin- coincide, because one can recover them from the extension at {oo} x M. Now the elements y\ must also coincide. This finishes the proof of Theorem 11.1. Remarks 11.7 (a) For most of the singularities there also exist metrics g other than those constructed in Theorem 11.1 such that the F-manifold (M, о, е, Е) together with g is a Frobenius manifold. Even the eigenvalues of Vs E are different. These other metrics arise in the same way as those in Theorem 11.1, but after changing a very fundamental choice: One replaces the natural order < on the set {a | e ~2nia is an eigenvalue of the monodromy } U Z by another order which satisfies G.9). 11.2 Deformed flat coordinates 205 Then the spaces H°°, C, V>~O0, and НЦ are still the same, but the V- filtration, the spectral numbers of Яд', and the corresponding filtrations F. and F#a'« from G.24) and G.27) usually change. Often opposite filtrations exist and the whole construction in this section can be carried out. (b) Let us sketch this for an example of M. Saito [SM3, 4.4]. The semi- quasihomogeneous singularity / = x$ + x\ + x^x\ has spectral numbers ot\ < ... < Q!25 which are as an unordered tuple (—1 + '+{+2 | 0 < i, j < 4). One chooses st € C"*' as in A1.14) and A1.15) (this includes the choice of an opposite filtration). It then turns out that v° — Sjfor/ > 2andu° = si+a-Va.S25 for some a e C\ with si e C~2'3 = C", УЭ;525 e C~x^ = C5. If one now chooses a new order -< as in G.9) with —5^-5 then the spectral numbers for this order are (no longer with indices fitting to their order) 5, = a, for 2 < i < 24 and ci{ = a25 - 1 = -5, «25 = <*i + 1 = \ Now one can choose si = s,- for 2 < i < 24and?i = Then v° = v? for 1 < i < 24 and v%5 =125- The whole construction in this section can be carried out. One obtains a Frobenius manifold with [5,-, E] = A + Si — a,)<5;. (c) The different choices of opposite filtrations and of orders ~< seem to be related both to certain transformations of Frobenius manifolds of Dubrovin [Du3, Appendix B] [Du4, first part of Lecture 4] and to the Schlesinger trans- transformations in [JMU][JM]. \- 11.2 Deformed flat coordinates In this section we want to establish series of functions с\^ е Ом,о which are implicit in the construction of section 11.1 and which are very close to Dubrovin's deformed flat coordinates. These deformed flat coordinates play a major role in many of Dubrovin's papers on Frobenius manifolds. They lead to rich hidden structures [Du3, Lecture 6], [DuZl], [DuZ2], [Gi8], whose meaning for singularities has still to be explored. We will use our series of functions in chapter 12 in order to establish a canonical complex structure on the /x-constant stratum and study a period map. Remark 11.8 The most detailed description of the deformed flat coordinates can be found in [Du4, Lecture 2]. We refrain from repeating that here and restrict ourselves to some comments. The first structure connection У(~^ (Definition 9.6) on the lift pr*TM of the tangent bundle to P1 x M is flat and has a logarithmic pole along {0} x M. The same holds for the induced connection on the dual bundle pr*T*M.
206 Frobenius manifolds for hypersurface singularities Dubrovin's deformed flat coordinates %(t, z) are multivalued functions on C* x M whose restrictions to slices {z} x M form locally flat coordinates with respect to V(~2>. They also have the best possible behaviour with respect to dz. To make that precise, let dM denote the differential only with respect to the coordinates on M, not with respect to z. Then the differentials &M% are multivalued flat sections in pr*T*M |C* x M with respect to the induced connection. The deformed flat coordinates Ц are usually written as a tuple , z) 7m(t, z)) = @i(r, z),.. •, em(t, я A1.42) [Du4, B.83)], [DuZl, B.63)], [DuZ2, C.13)]. The matrices /i and R are sup- supposed to be chosen such that the following holds: The functions 0,-(f, z) are holomorphic in С* х М. Their differentials d^fy form a basis of elementary sections in pr*T*M |C* x M. (Here ±д is the matrix for the (by assumption) semisimple residue endomorphism, the matrix R carries the Jordan block structure of the monodromy.) The coefficients 9jp(t) in the expansions Oj(t, z) = J2p>o^ip^z>> are more or less equivalent to the coefficients c,y+ below. 'More or less', because our coefficients come from the second structure connection V(~5> which is close to, but in general not equal to the Fourier dual of V(~5> (cf. Remark 9.9 (b)). The 9jP are the real starting point for the profound structures in [Du3, Lecture 6], [DuZl], [DuZ2], Gi8 We consider the same situation as in section 11.1 and use the notions es- established there. Most important are the global sections u,- and the elementary sections si. The germs in H@>@0|0) of the global sections v, will be developed in series 1<У</г p>0 with coefficients c]f € Ом,о- Because of A1.12) and A1.14) one has A1.44) where? = К Therefore the germs v,- in (oo, 0) can be written uniquely as sums of elementary sections ~zpSj. But we want to write them with (—Vs. )pSj. For a $ Z<o there is no problem. The sheaf W' has a logarithmic pole along {oo} x M, and — dz = ?(?%). Therefore the sections (—Vaz)pSj with ay — p = a and p > 0 form a basis of the same subspace of С as the sections "zpSj. But for а е Z<o there are two related problems: 77.2 Deformed flat coordinates 207 (I) The sections (—V3.)''sy with a,- — p = a and p > 0 generate a strictly smaller subspace of С than the sections ~zpSj. A priori it is not clear that the elementary part of d,- in Ca is in this smaller subspace. But it will turn out to be the case. (II) (,-^b.ysj — 0 for certain j and p. The coefficients с(У in A1.43) for these j and p are not unique, but arbitrary. In order to obtain unique coefficients с? we have to introduce a dummy expo- exponent ? e N and consider the analogon of A1.43) for (—Va. )^ ы for all ? e N. That shifts the orders of the elementary sections out of the bad domain Z<0. Each coefficient c\f will be uniquely defined by the expansion of (-V3.)~^u, for sufficiently large ?. The same trick will be used to establish the relations between the c\j\ It is clumsy, but worth the effort. With the first structure connections, one would not have this difficulty. Of course, (—V3z)~^D; is the section in W@'|C x M which is welldefined by the proof of Theorem 10.7. The germ at (oo,0) is a germ in Ц~$ ¦ Theorem 11.9 (a) Let u,- and s-, be as in section 11.1. There exist unique coefficients cff e Ом,о {p € Z, i, j e {1,..., fi}) such that the germs at (oo, 0) satisfy (-У3Г^= Y.Y.cll?(-Va,)"-llsj (П.45) Om,q be A1.46) A1.47) for all ? e N. The coefficients satisfy c]f = Ofor p < 0 and cfj} = 50 (b) Let Si = ir'(i>i) be the flat vector fields from A1.32). Let a\} € the multiplication coefficients with Si о Sj = ^^ af-S^. Then and C Especially Si = e, akjx = <5,ь and «,#=4. A1.48) A1.49) A1.50) The last equation says that r,- := c(,)' are flat coordinates on M with -^ = 5,-.
208 Frobenius manifolds for hypersurface singularities (c) For any i,je[l,...,n},re Z, s,-M+waOr= ? B-^M p,q: p+q=r к The right hand side has for r > 1 the linear terms (—l for r > 2 additional quadratic terms. Especially С с; V C(D (d) Let nij eCbe the coefficients with (cf. A1.6)) Then nij = Oforoij — 1 — а,- ф 0 and Z-Vi = (a,- + l)V^'v; +^W,7V; ¦ and for p > 2 1_у + с;д+1_,. anrf A1.52) A1.53) a,)t»y, A1.54) (П-55) 77ге Euler field E satisfies EoSi = J2n>JSJ ~ ?cU){aJ E = A1.58) Proof, (a) Because of u,- = v| (V^*v,), the elementary part of v,- in C" is contained in v| Ca+$ for any | e N. The same holds for the elementary parts of (—V8i)~?v,. This solves the problem (I) before Theorem 11.9. For each f one has an expansion A1.45). The coefficients c)j are independent of ? and uniquely determined by the expansions for sufficiently large ?. Now с? = О for p < 0 is trivial and cf^ = 5y follows from A1.18). (b) Formula A1.46) is A1.33) rewritten. One complements (—Vg7)~? on both sides and inserts the expansions A1.45). One obtains A1.47) by comparison of the coefficients of (— V3j)p~Ss, for sufficiently large f. The rest is obvious. 11.2 Deformed flat coordinates 209 (c) First we formulate an approach which does not work, but which helps to understand A1.51). One would like to ignore that certain (— Va )pSk vanish. One would like to extend the pairing Kf from section 10.6 Ом off ЭГ1}} [3z]-sesquilinear and calculate Kp(Vi, Vj) with this extension of Kf and with the expansions A1.45). One would obtain A1.51) because of A1.15) and A1.26). The real proof is somewhat similar to the proof of A1.26). One has to con- consider the right hand side of A1.51) as a power series in OMfi and discuss the coefficients separately. Let Xb ..., Xm e TM,o (the empty set for m = 0), r > 0 and f > r + m. Then the number 8mO8ill+i-j8ro ={Xl...Xm 4r?)(( -V,,)*^, ( -V3ir2S)))@) A1.59) can be calculated as follows: One inserts the expansion A1.45), uses Theorem 10.11 (v), Theorem 10.28 (i), A1.15), and the properties of Kf. One obtains the value at 0 of the derivative by Xi... Xm of the right hand side of A1.51). Details are left to the reader. (d) Formulas A1.54) and A1.55) will be proved simultaneously by the follow- following calculation. The first step uses A1.45), A1.53) and [г, (-Эг)*] = *(-Эг)*-'. In the second step the expansion A1.45) is applied to the terms for p = 0, in the third step to a part of the terms for p — 1. j.p j.k p>\ ? ? (c
210 Frobenius manifolds for hypersurface singularities -p- к р>2 The germ at (oo, 0) of z • v,- is contained in V (п.60) A1.61) The restriction toCxM of г-v,-is containedin7^(°>|CxAf. Therefore z-n; is con- contained in ф;. Om,o ¦ Vj Фф; Одт.о • Va~' vj. The difference between (-V8. )~*z¦ v, and the first two lines of A1.60) is an element of (-У3.Г*(ф; Ом~о ¦ vj ф ф;- О мл ¦ Vg^ Vj). If it were nonvanishing then the expansion into elementary parts would yield a nonvanishing part in (-^Зг)~?(ф. Ом,о ¦ Sj Ф ф ¦ Ом,о • V^sj). But the sum of the last three lines of A1.60) does not have such a part. Therefore the difference and this sum are both vanishing. This gives A1.54) and A1.55). (e) Formulas A1.56) and A1.57) follow from A1.39) and A1.54). Again with A1.39) one obtains a,v; = V?+z3jv,. One puts into both sides the expan- expansion A1.45) and applies A1.53) on the right. Comparison of coefficients gives A1.58). n Remarks 11.10 (a) There is a rich structure hidden in the equations in Theorem 11.9, as is clear from Dubrovin's and Zhang's work. The most impor- important coefficients are the c(,f in view of A1.48). Central for Dubrovin is equa- equation A1.47), usually written as a second order equations for the c<f+1). For the /г-constant stratum we will mainly regard the equations in (d). Equation A1.55) shows that the coefficients c\]j> and the c\f with otj - p - a,- = 0 determine all coefficients c^\ (b) For fixed coefficients c\f e С with c\f = 0 for a, - p - щ < 0 the equations A1.55) and A1.52) imply A1.51). This was proved in [He4, Proposition 5.5]. It was used to describe coordinates for the fibres of the clas- classifying space DBL for Brieskorn lattices as a bundle over a classifying space Dpmhs for polarized mixed Hodge structures. But if only c\f -OfoTUj-p-at < 0 then the coefficients with aj -p -a,- = 0 satisfy equations of the type A1.51) which do not follow from A1 55) and A1.52). 11.3 Remarks on mirror symmetry 211 11.3 Remarks on mirror symmetry Dubrovin's definition [Dul][Du3] of Frobenius manifolds formalized a part of the structures which the physicists Witten [Wit], Dijkgraaf, E. Verlinde, and H. Verlinde [DVV] had found studying (moduli spaces of) topological field theories. Different ways in physics to establish such structures lead to phenomena which are now comprised in the famous notion mirror symmetry. There are several very different versions of mirror symmetry. A vague para- paraphrase of one version is that one has two sides, the A-side and the B-side: on the A-side one looks at data related to the Kahler geometry of a manifold X, on the B-side one looks at data related to the complex geometry of a family of manifolds Y,. These data should be isomorphic if the manifold on the A-side and the family of manifolds on the B-side are mirror dual to one another. See [Vo] for a detailed discussion and references. The structure on the A-side of this version of mirror symmetry comes from genus 0 Gromov-Witten invariants and is called quantum cohomology . Within the frame of algebraic geometry it has been established by Kontsevich, Manin, and many others (see [KM], [Man2], and references therein). If X С VN is a manifold with Hodd(X, C) = 0, then its quantum cohomology can be encoded in a formal germ of a Frobenius manifold, where the formal germ of a manifold is (H*(X, C), 0). The part of the structure which lives in some sense on (H2(X, C), 0) С (Н*(Х, С), 0) is called small quantum cohomology (see [Man2] for any details). The structure on the B-side of this version of mirror symmetry is related to period integrals, Picard-Fuchs equations, hypergeometric functions. If Y, is a family mirror dual to X, then the parameter t should be in a space isomorphic to a neighbourhood of 0 in H2{X, C), and the data from Y, should be isomorphic to the small quantum cohomology of X. The firstexample of this was proposed be Candelas, de la Ossa, Green, Parkes [CDGP]. It was proved together with many other cases by Givental [Gi4]. In view of the full quantum cohomology and the Frobenius manifold on the A-side, one may ask about an extension of the family Y, and the data on the B-side, such that in that case one also obtains a Frobenius manifold, and such that it is isomorphic to the Frobenius manifold on the A-side. That would be a stronger version of mirror symmetry. One approach to this is provided by the Barannikov-Kontsevich construc- construction [BaK][Bal][Ba2]. Using tools from formal deformation theory, there an extended moduli space M (a priori a formal germ, in good cases a man- manifold) is constructed such that a subspace Mcs (cs for complex structure) parameterizes a family Y, of complex manifolds. Under certain assumptions
212 Frobenius manifolds for hypersurface singularities О w the space M. is equipped with the structure of a Frobenius manifold. Build- Building on Givental's results, Barannikov proved isomorphy of this Frobenius manifold with that from quantum cohomology of X when X is a projec- tive complete intersection Calabi-Yau manifold [Bal]. A major and still unsolved problem is to find objects which are parameterized by points t e M - Mcs. Another approach to get Frobenius manifolds on the B-side is motivated by a proposal of Givental [Gi3] and Eguchi, Hori, Xiong [EHX]: if X on the A-side is not a Calabi-Yau manifold, but, for example, a Fano manifold, one should consider on the B-side as mirror dual a family (У,, /,), where /, : Yt —> С is a function on Y,. If X is Calabi-Yau then /, should be constant, but in other cases it may be a function with isolated singularities on the (now not compact, but usually affine) manifold Yt. Then the data on the B-side for the family (Y,, /,) come from oscillating integrals. Givental proved in many cases that they are equivalent to the small quantum cohomology of the mirror dual manifold X [Gi3][Gi4][Gi5][Gi6]. In many cases the parameter space of the family is isomorphic to an open domain in H2(X, C), and at all isolated singularities of the single functions /(, the family is a /x-constant deformation. Now the extension of the family (Yt, /,) to a family with parameter space M with dim M = dim H*(X, C) is often not difficult: one has to consider also deformations of the functions /, which are not д-constant at the singular points. One arrives at a situation which generalizes the semiuniversal unfolding of a local singularity / : (C"+1, 0) -* (C, 0). Sabbah saw that one can often establish the structure of a Frobenius man- manifold by a procedure which is similar to that in section 11.1. But one needs oscillating integrals instead of the GauB-Manin connection. This is because the middle cohomology group of the smooth fibres f,~i(z) usually has a di- dimension Ф dim M. Also, the results on the existence of a Frobenius manifold are not as complete as in the case of a local singularity. This is discussed in [Sab2][Sab4]. The best positive results are due to him [Sab3][NS]. We give a sketch of our understanding of the situation and part of his work in the next section. The family of functions mirror dual to P" was proposed and studied in [Gi4] and [EHX]. The construction of its Frobenius manifold was carried out by Barannikov [Ba3]. It fits into the frame given by Sabbah. 11.4 Remarks on oscillating integrals The version of mirror symmetry which was discussed in section 11.3 asks in its strongest form whether certain Frobenius manifolds from quantum cohomology 11.4 Remarks on oscillating integrals 213 of projective manifolds are isomorphic to Frobenius manifolds coming from suitable families of functions on suitable manifolds. Semiuniversal unfoldings of local singularities are not suitable families of functions, simply because their spectral numbers never fit to those from quantum cohomology. But one may hope to find suitable families of functions starting from func- functions which were studied by Sabbah [Sab2][NS]: M-tame functions on affine manifolds (definition at the end of this section). Sabbah indicated how one may proceed to construct Frobenius manifolds from them [Sab3][Sab4]. The construction is similar to that in section 11.1. But instead of the GauB- Manin connection and a second structure connection one considers oscillating integrals and a first structure connection. The two most difficult steps are: (a) to construct a meromorphic connection on a trivial vector bundle over P1 x M which looks like a first structure connection; (b) to find a certain global section (a. primitive form) which induces an iso- isomorphism between the lifted tangent bundle n*TM and this trivial bundle. Step (a) involves (the spirit of) oscillating integrals and the choice of an opposite filtration. By some work of Sabbah, it can be carried out for a large class of functions, all M-tame functions on affine manifolds. But the existence of a primitive form in step (b) is for the moment only clear for a much smaller class of functions, for polynomials with nondegenerate Newton boundary. In [Ba3], Barannikov treated (independently) the family of functions mirror dual to Pn+1. Although the presentation is completely different, this fits into Sabbah's frame. Not all steps in Sabbah's programme have been worked out in detail. And we will not do it here. We will restrict ourselves to formulate this programme in 5 steps and then comment upon some aspects of the steps. Step 1: Construct a family of functions /, : X, -> A, t € M С О\ with the following properties. (a) The set A = {z e С | \z\ < »?} is a disc, X, С CN is a Stein manifold of dimension и + 1, X := \Jt X, x {t} С CN x M is a manifold, M is a (small) open subset of C*. (b) The function/, : X, -* Д has isolated singularities with д = Z^eSingC/,) m(/(. x), and for any 7 e M the restriction of the family of functions to the multigerm at Sing(fy) x {?} С X is a product of semiuniversal unfoldings of the singularities of jj. (c) The map <p : X -*¦ A x M (jc, О м- (/,(*), 0 A1.62)
214 Frobenius manifolds for hypersurface singularities 11.4 Remarks on oscillating integrals 215 fir. is a C00-locally trivial fibration outside of a discriminant V С А x M. For any (г, t) e V, the fibre <p~l(z, t) = /,~"'(z) x {t} is singular, and there exist arbitrarily small neighbourhoods U\ С X of Sing(cp~l(z, t)), (/2CA of г, and U-i С Moffsuchthatthe restriction of ipto^~'((/2 x (/з)-(/[ is a C°°-locally trivial fibration. Step 2: Construct a flat bundle Ни/ of rank д over С* х М (definition in A1.66)) with connection VLef such that the fibres of the dual bundle can be interpreted as spaces of Lefschetz thimbles. Let Иье/ be the sheaf of holo- morphic sections of #z.e/. Define a certain pairing on the sheaf n^Tiuf (here я : С* x M ->¦ M). Step 3: Extend the sheaf TiLef to a free 0CxM-module 1iflf such that the pair CW^e/» VLef) has a pole of Poincare rank 1 along {0} x M. This uses the GauB-Manin connection of <p and a Fourier-Laplace transformation with parameters. A geometric interpretation imitates oscillating integrals on a (co)homological level. Step 4: Try to extend the sheaf fiflf to a free Or\ xW-module Hflf with a logarithmic pole along {oo} x M. This is aBirkhoff problem with parameters. Sabbah showed that this is solvable for similar reasons as in the case of a local singularity if one starts with an M-tame function on an affine manifold. He established a mixed Hodge structure on the space Hfy of manyvalued flat global sections of Huj [NS]. An opposite filtration to its Hodge filtration induces a solution of the Birkhoff problem. Step 5: Try to find a global section V\ (a primitive form) in Hflf which is flat at {oo} x M with respect to the residual connection there, which is an eigenvector of the residue endomorphism at {oo} x M, and which induces an isomorphism Тм —*¦ л^И^ by a period map. Then a part of the pairing in step 2 induces a flat metric in TM such that M gets a Frobenius manifold. This isomorphism also identifies (Ti-flf, VLef) and the first structure connection (V(~5~!), рг*Тм) (with the coordinate г = — J-, whereЛ is a coordinate on C* used below). Because of condition (b) in step 1, M will be a massive F-manifold with Euler field and smooth analytic spectrum. The cohomology bundle of the fibration <p from A1.62) is Hn:= У Hn(fr\z)X). A1.63) (г,/)еДхЛ/-Й If the spaces X, are not contractible then in general the fibres of Hn do not have dimension д. In that case there is no chance of extending the GauB-Manin connection to something isomorphic to a second structure connection. But by condition (c) in step 1, the spaces of Lefschetz thimbles have the right dimension д. This is the reason why oscillating integrals will be used. To define Lefschetz thimbles, fix for a moment# e C* and t e M. For sufficiently small ? > 0 let Д<Л) := {z € A | \z — г]щ\ < e} be a neighbourhood in Д of the boundary point т)щ. Then X\K) := f~l(A{n)) is homotopy equivalent to a smooth fibre off. A Lefschetz thimble in X, with boundary in x\ is an (n + l)-cycle Г = Цехао,]]M(г)> where У : [°> !] -> л is a Path with K@) 6 f(Sing(f,)), K(@, D) П f,{Sing{f)) = 0, K(D 6 A№), and S(z) for z e y([0, 1]) is a continuous family of n-cycles in /r~'(z) which vanish for z —»¦ y@). Condition (c) in step 1 and standard arguments (e.g. [AGV2, ch. 2], [Lo2, E.11)]) show that one obtains a space homotopy equivalent to X, if one glues ix Lefschetz thimbles to xf\ (First one chooses t e M such that f has д A [-singularities with different critical values and considers Lefschetz thimbles over ц. nonintersecting paths from f(Sing(f,)) to Д<л).) Therefore « A1.64) and this space is generated by the relative homology classes of Lefschetz thim- thimbles. There is an exact sequence 0 -+ Hn+l(X,, Z) -> Hn+l(Xt, Xf\ Z) A1.65) ^ Hn(X?\l)-* Hn(X,,Z) ^ 0. The bundle Hief in step 2 is HLef.= U Нп+1(Х„Х?\СУ. A1.66) (Я,;)еС*хМ A pairing on ж» 7^/ can be defined using the intersection form for Lefschetz thimbles, which is a perfect pairing < .,. >: Hn+1(Xt, Xf\ Z) x Н„+1(Х„ Х\-к\ Z) -> Z. A1.67) It induces a pairing <.,.>* on the dual spaces. Then KLef : n^Hief X ntHief -»¦ Л*ОсхМ (а,Ь)\-*(Н\-><а(П),Ь(-П)>*) A1.68) is Ом -bilinear and sesquilinear тй. The definition of the extension Tlflf of H-uf to С х М in step 3 is not easy. One has to start with an extension 7Y@) to Д x M of the sheaf of sections of the cohomology bundle Hn. This should be defined using relative (n + l)-forms as
ел; 216 Frobenius manifolds for hypersurface singularities in chapter 10, but the definition in Lemma 10.2 probably gives the right object also in this more general situation. Then a Fourier-Laplace transformation with parameters (i.e. keeping t and 3,, changing z ь> —3r, dz ь> r, where r := |) should give H(°lf. The best way to handle this has still to be fixed. A geometric interpretation looks roughly as follows. One extends the homology bundle and the cohomol- ogy bundle H" of cp to flat bundles over С x M - T>. If Г = LUy([o,i])S^ is a Lefschetz thimble as above, one extends the family [S(z)] of homology classes of cycles to a continuous family in the extended homology bundle over y([0, oo)), where y([l, oo)) is the half line which starts at y(l) and goes in the direction -щ. If со е 7r*7^@) now extends to a global section over CxM with moderate growth along {oo} x M, one defines [е-г*а>](\Г]):= [ y([0,oo)) n co{[8{z)\)iz. A1.69) This imitates an oscillating integral. One has to show that it gives a holomorphic section [e~z^co] in HLef. Then such sections generate Hflf as an OcxM-module As soon as one has proved formulas which establish the Fourier-Laplace correspondence between sections of 7^@) and H\2/ (i-e- г и- -Эг,Эг к t, where т = |), it is not hard to see that T~L\jef is a free Осхм-module with a pole of Poincare rank 1 along {0} x M. If the family /, : Xt —*¦ Л is a semiuniversal unfolding of a local singularity then (—1);lt~ BgL+i Kief on itt'Hflf is the Fourier dual of K. Saito's higher residue pairings Kp on 7r*7^<0) from section 10.4. This follows from [Ph5,2eme partie 4]. In order to find an extension to a free Op \xM -module T-6?lf with a logarithmic pole along {oo} x M, one can restrict to a fixed parameter t e M. Let n"LL\t be the free Ос-module of restrictions of sections in Tiflf to Hief\cx(i)- It has a pole of order < 2 at 0. One wants to solve the Birkhoff problem for T^ilffc in the spirit of sections 7.2 and 7.4, by defining a filtration F* in H?°f from 7^j^|7 and choosing an opposite filtration. But the procedures in sections 7.2 and 7.4 do not apply, because H*j%f \r has a pole of order < 2 at 0. Sabbah found a very nice variation of these procedures. One has to look at global sections in H^ \j with moderate growth at oo and has to define principal parts for them using the V-filtration at oo. The principal parts induce a filtration F' on Я^-. Any opposite filtration gives rise to a solution of the Birkhoff problem for nfef \j. See [Sab4, III 2.b and IV 5.b] and [Sab2] for details. 11.4 Remarks on oscillating integrals 217 A profound result of Sabbah is that F* is a Hodge filtration of a mixed Hodge structure if f~t : X~t —> Д is the restriction of an M-tame function /7 : Y —> С on an affine manifold К DI7 [NS][Sab2]. Then opposite filtrations exist and the Birkhoff problem is solvable. Definition 11.11 ([NZ][NS]) The function /7 : Y -+ С is M-tame if for some closed embedding Y с <CN of the affine manifold Y the following holds: for any r] > 0 there exists an R(n) > 0 such that the spheres S™?1 = {x € CN | \x\ = R(r])} are transversal to ffl(z) if г € Д. Then the restriction /7 : B^|' П ffl(A) -*¦ Д is the analogon of a Milnor fibration of a local singularity, if Д contains all critical values of /7 : Y -»• C. In [Sab2] a related more algebraic notion, cohomologically tame, is considered. But it is not clear whether it implies M-tame. Suppose now that /7 : Y -> С is M-tame. Then one can realize the definition of W^fr and the formula A1.69) by proper oscillating integrals over Lefschetz thimbles in the sense of [Ph4] [Ph5]. One can also realize the family /, : X, -»• Д with t € M by global functions /, : Y -> С But in general these functions will not be tame in any sense and will have nonisolated singularities or too many isolated singularities. In order to obtain a family of functions as in step 2, in general one has to cut out bounded spaces X, С Y, which contain the singularities of ft, but exclude the bad phenomena coming from infinity for t Ф t. In step 5, the existence of a primitive form is unfortunately at the moment only clear for a much smaller class of functions, for polynomials /7 : Cn+1 —> С with nondegenerate Newton boundary [Sab3][Sab4]. The problem is that the analogon of M. Saito's Theorem 10.33, that the smallest spectral number «i has multiplicity 1 and corresponds to volume forms, is not clear in the case of globalfunctions f?: Y -> C. But it is not difficult to establish this in concrete cases, and it may well hold for many classes of functions important for mirror symmetry.
о Chapter 12 д-constant stratum Two applications of the results of chapter 11 are given here. In section 12.1 a canonical complex stracture on the д-constant stratum of a singularity is es- established. The most difficult part is to prove that it is independent of the choice (U., у\) in Theorem 11.1. In section 12.2 the period map from ад-constant stra- stratum to a classifying space for Brieskorn lattices from [Hel][He2][He3][He4] is taken up. It is shown that the map is an embedding. This strengthens a result of M. Saito. 12.1 Canonical complex structure The д-constant stratum of a singularity can be equipped with a canonical com- complex structure. This will follow from the construction of Frobenius manifolds in chapter 11 and from Varchenko's result that the spectral numbers are constant within a д-constant family (Theorem 10.32, [Va3]). One can write down the complex stracture quite explicitly for a choice of an opposite filtration. But the proof that it is independent of this choice is technical. Let / : (C"+1,0) -* (C, 0) be a holomorphic function germ with an isolated singularity at 0, with Milnor number д, and with n > 1. We choose a representa- representative F : X -*¦ Д of a semiuniversal unfolding F: (C"+1 x CM, 0) -*¦ (C, 0) as in section 10.1, that means, Д = B\ сС,М = В,"сС',Л'= F-\A)n(B^+i x M) for s, S, 9 > 0 sufficiently small. The function F, : X П {Bne+X x {t}) -»• Д is the restrictionto the parameter t = (t\,..., fM) € M. The д-constant stratum 5M с М is S^={t €M \ Ft has a singularity with Milnor number д and critical value 0}. A2.1) It is a subset of the discriminant V С М, V = {t e M | F, has a singularity with critical value 0}. A2.2) 218 12.1 Canonical complex structure 219 Both can be described using the canonical stracture of M as an F-manifold {M, о, е, Е) (see section 10.1 and section 5.1 for the definition of this structure). The discriminant is V = det(W)~'@), where U is the multiplication with the Euler field. Gabrielov [Ga], Lazzeri [La], and Le [Le] proved that F, has only one singularity if all its singularities are concentrated in one fibre. This is used in the following characterization of 5Д. Theorem 12.1 (a) The ^-constant stratum 5M С М is an analytic subvariety. It is 5Д = V П [t € M | (T,M, o, e) is irreducible} A2.3) = {teM\U: T,M -»¦ T,M is nilpotent} A2.4) = {t € V | the e-orbit through t intersectsV only in t]. A2.5) (b) The germ ((M, 0), o, e, E) of an F-manifold together with the germs (U, 0) and EM, 0) is independent of the choice ofF, it is unique up to canonical isomorphism. Proof, (a) Equation A2.3) holds because (T,M, o, e) is isomorphic to the sum of the Jacobi algebras of the singularities of F,. Both sets on the right are analytic (cf. Proposition 2.5 for the second). So SM is analytic. The endomorphism U corresponds to the multiplication by F, on the Jacobi algebras. Equations A2.4) and A2.5) follow with the result of Gabrielov, Lazzeri and Le. (b) Theorem 5.4. ? Now we consider the construction in chapter 11. We will quite freely make use of the notations introduced there. The choice of an opposite filtration U. and a vector y\ as in Theorem 11.1 and the choice of s\,..., s^ as in A1.14)- A1.16) induces a flat metric g on M, a basis 5i,..., 5M of flat vector fields, and coefficients c\f (Theorem 11.9). Theorem 12.2 Let (U., y\), s\, ¦ ¦ ¦, V S\, ¦ ¦ • > V andc\f be as in chapter 11. Define coefficients e,y 6 Ом,о by E о &, = ]TV SijSj. Then there is an equality of ideals (ey I a} - 1 - a,- < 0) = (c<]> | aj - 1 - a, < 0) A2.6) = (cff \otj-p- a,- < 0) С Om,o. This ideal defines the /г-constant stratum EM, 0) С (М, 0).
220 fi-constant stratum Proof. In A1.56) nij = 0 for otj - 1 - а,- ф 0. This shows s,j = -(a,- - 1 - oti)c]V for ay - 1 - a,- ^ 0 and the first equality. The second equality follows with A1.55). We have to show M = [t 6 M | c\)\t) = 0forotj-\-a,- < 0}. A2.7) The matrix (e,j(r)) for r as on the right hand side is nilpotent. Equation A2.4) gives the inclusion Э . For the opposite inclusion С we need Varchenko's result that the spectral numbers are constant within a д-constant family (Theorem 10.32, [Va3]). Fix a parameter t e S^. The restriction to (Д x {r}, @, t)) of the sections of TC^ gives the Brieskorn lattice H([(t) of the unique singularity of F, (see Remark 12.3). Its spectral numbers are ab ..., aM by Theorem 10.32. Let v\ € Hg(t) be the restriction of the global section u,- to this Brieskorn lattice. Consider the expansion A1.45) for v't. It is finite, c^\t) = 0 for a,- - p < c*i, because of Я^'@ С V". The inclusion С in A2.7) is equivalent to the claim v? 6 V"'. This holds obviously for v i. If any v\ with i > 2 had a principal part of order < otj then this would be independent of the principal part of v\ because that contains s\ and v\ with i > 2 does not contain s\. The second spectral number would be smaller than аг, a contradiction. Therefore v\ e V1 for i > 2. The claim follows inductively. D Remarks 12.3 The identification of the restrictions to (Д x {t}, @, r)) of the sections of Tim with the sections in the Brieskorn lattice Щ'A) of F, for t e S^ is not as self-evident as it looks. The restriction to (Д x {t}, @, t)) of the fibration ip : X -*¦ A x M, (x,t)i-*- (F(x, t), t) of the unfolding F is not necessarily a Milnor fibration for the unique singularity of F,. It could happen that one has to choose e(t) <? s to make sure that all balls in C"+1 which have radius s' < e(r) and are centred at the singular point of ^,"'@) intersect F~l(Q) transversally. But according to [LeR, Lemma2.2] the restriction cp\(A x {t}, @, r)) is fibre- wise homotopy equivalent to a Milnor fibration for F,. The cohomology bundles are canonically isomorphic. The restrictions to (Д x {t}, @, t)) of sections of Нф) give the Brieskorn lattice ЩЦ) of Ft. Also, the space of multivalued flat global sections of the cohomology bundle of F, is canonically isomorphic to H°°. The spaces of elementary sections for this cohomology bundle can be identified with the spaces Ca of elementary sections near {oo} x M for the extended cohomology bundle of <p. Compare the observations A)-(Ш) before Lemma 11.2. 72.7 Canonical complex structure 221 This allows the Brieskorn lattices H^t) for all t e 5M to be considered as sublattices of the same lattice V >-1. In section 12.2 we will consider the period map SM ->• DBL to a classifying space for such lattices. Theorem 12.4 The ideal in A2.6) is independent of the choices of su ..., sM and (U.,y\).It defines a canonical complex structure on the [i-constant stratum Proof: 1st step: We fix an opposite filtration U.. That does not induce a unique metric, but a unique flat structure on M. Then adE acts semisimple on the space of flat vector fields with eigenvalues -A + ai - ctj). Let 5i,..., 8^ be any basis of flat vector fields with [5,, E] — A +c*i - a,-M,-. Define?,^ € OMfi by E о % = J\ еи ¦ 1j. Then the ideal (eu \ ot} - 1 - a,- < 0) is the same for all such choices of 1\,..., \, because a base change mixes only % with the same orders a, and ay. Therefore the ideal in A2.6) depends at most on (/., not on y\,s\, ...,jm. 2nd step: Let U. and U. be two opposite nitrations. They induce subspaces G" and Ga of Ca as in A1.5). They satisfy and vice versa. Choose sections ji, ... ,5^ as in A1.14) and A1.15) for U. There exist unique sections 5,- 6 G°" of the form Si= with df = Sij. They satisfy A1.15). Let v,- and U, be the corresponding global sections for the two extensions of Hi0) by U. and U.. Over С х М they are bases of the same sheaf, along (oo)xM they generate different exten- extensions with logarithmic pole along {00} x M, but both with spectral numbers -ai,..., —aK € (-n, 1). Therefore the coefficients in й = фв,,(г,О-«/ 02.9) satisfy A2.10) m=0
X Xr' 222 \i-constant stratum Because of A1.54) one can rewrite this with Vr1, A2.11) with unique ^т) 6 dM. Now one can calculate the coefficients c^ in terms of the c-?\ One uses A1.45) for и,- and for u,-, A2.11) and A2.8). One obtains l jtk,m,p,q: ij Cjk "kl A2.12) Here note that b^ = 0 for m < -n and for m > 0; dff = 0 for q > 0 and for _. _ _. / (\. j\") __ a . \p) f\ i?_ a @) p "/ "~ 4 ~ ak r U, "W = "kit Cjk = U 1ОГ /J < (J; Сд ^ од. We want to show c^ e (сд' | c^ - p - aj < 0) for a/ - r - a,- < 0. Because of A2.12) it is sufficient to show the following claim: bV 6 (c<fk I ak ~ P ~ <*j < 0) fora;-r-a,- < 0. A2.13) Claim A2.13) will be proved inductively over the size of щ — r - a,. Fix fi < 0 and suppose that A2.13) is true for a, - r - a,- < p. Fix a triple (i, /, r) with a; - r - a,- = 0 and r < 0. Then c^' = 0. Equation A2.12) gives a relation for the monomials on the right hand side. Modulo the ideal (c^ | ctk — p — ctj < 0) the only monomials left are those with p > 0, ak - p - oij = 0, otj -m-a; = and those with A2.14) A2.15) Because of A1.24) c)pk\<X) = 0 for p > 0, ak - p - a,- = 0. /2.7 Canonical complex structure 223 So, A2.12) modulo the ideal is a linear relation between ?>-m) with aj ~m — a, = P; the only eventually invertible coefficients are 1 for bff and d^~m) for fcjj' with m > r. Putting together the relations A2.12) for all triples (i,l,r) as above, the coefficient matrix is triangular and unipotent for t = 0. Therefore the claim holds for (i, I, r) as above. This finishes the induction and the whole proof. ? The Kodaira-Spencer map yields a canonical isomorphism (cf. A0.1)) Э [dF I a0 : T0M -> Oc»+,,o/7/, — ^ — A2.16) °ti I °fi (=0J between the tangent space at 0 and the Jacobi algebra O/Jf. The space л A2.17) (cf. A0.87)) is a free module of rank 1 over the Jacobi algebra. Also, the V-filtration on Щ induces the V-filtration with VSif := pr(V П Щ) on af. Varchenko and Chmuto v [ChV] considered the tangent cone to the /U.-constant stratum. It is estimated by the space V\O/Jf) := {g 6 O/Jf | g ¦ Уа% С Va+1Qf}. A2.18) In fact, they took not only the У-filtration into account, but also some weight filtration. Some remarks on that will come in section 12.2. Here we will see that the space Vx(O/Jf) fits very well to the canonical complex structure on Lemma 12.5 The space VX{O/ Jf) С O/Jf is canonically isomorphic to the Zariski tangent space in TqM of the /л-constant stratum (S^, 0) with the canon- canonical complex structure from A2.6). Proof. Choose (?/., y\) an^ "l. ¦ • ¦ < vij. as in section 11.1. One arrives at the following identity in Я/ for X 6 TM,o for example through A1.35), A1.33), A1.49) i )l,«o •["*]• A2Л9) Because of Vuf = @a.>a C[u?], the space V\O/Jf) is isomorphic (viaa0) to {X|,_o еГ„М|Хе TM>0, X(cf/) \t=0 = 0 for aj-l- a,- < 0}. A2.20)
•г' 224 ^-constant stratum This is the Zariski tangent space of EM,0) with its canonical complex structure. D Remarks 12.6 (a) Choose (U,, y\) as in section 11.1. The functions r,- := cu are flat coordinates on M with <5,- = -^ (Theorem 11.9 (b)). The д-constant stratum is contained in the flat subspace of M {t € M | Xj(t) = OfOTdj -1-Cti < 0}. A2.21) In general this is only a coarse estimate. But for a quasihomogeneous singularity / this flat subspace is the д-constant stratum. In that case E\,-o = 0, and E = X^O+ai ~ctj)Tj8j by A1.57). All coefficients c\p are weighted homogeneous with weights or,- + p — aj because of A1.58). Therefore then - 1 - or, < 0) = (ту I ay - 1 - a, < 0) С >M,0- It coincides also with the ideal (tj | a,- — 1 — ori < 0). Varchenko [Va4] was the first person to see that this last ideal gives the д-constant stratum. (b) In the curve case (n = 1) the д-constant stratum is smooth ([Wah], see also [Matt]). Unfortunately, it is not clear whether the canonical complex structure is then the reduced structure. Results on the Zariski tangent space in [ChV] indicate that this should be the case for (generic?) curve singularities with nondegenerate Newton boundary. (c) For л > 2 the д-constant stratum can have singularities [Lu][Stv]. The construction in [Lu] is elegant, but it makes use of a part of the Zariski con- conjecture, that the multiplicity should be constant along the д-constant stratum. At the time of [Lu] a proof of it was announced which later turned out to be wrong. So [Lu] and an also elegant generalization in [VaS] concern the sets S/j, П {multiplicity constant}. In [Stv] it is proved for one example of [Lu] that this (singular) subset is at least a component of 5M, if it is not the whole space 5M. Results in [VaS] indicate that this subset (and probably 5M itself) can have arbitrary singularities (but the Milnor numbers in examples in [VaS] become astronomically high). (d) One may also expect that the canonical complex structure of 5M can be nonreduced. But examples are not known. Computing this canonical structure for higher Milnor numbers is difficult. 12.2 Period map and infinitesimal Torelli Let / be a singularity as in section 12.1 with semiuniversal unfolding F, base space M, and ^-constant stratum 5Д С М. We argued in Remark 12.3 that 12.2 Period map and infinitesimal Torelli 225 the Brieskorn lattices H^t) for all t 6 5K can be considered as sublattices of the same lattice V>~i. Also the Hodge nitrations F'(t) can be considered as filtrations on the same space Я°°. By Varchenko's result (Theorem 10.32, [Va3]) that the spectral pairs are constant within the jx-constant stratum, these Hodge filtrations are contained in the same classifying space Dpmhs for polarized mixed Hodge structures. By definition, the classifying space DBL from A0.94) for Brieskorn lat- lattices is a fibre bundle over DPMHS, with projection prPMHS : DBL -» DPMHS. Therefore we obtain two period maps Ф : DBL, г н> Яо'@, A2.22) DPMHS, t h+ F'(t). A2.23) There is also a global period map. In section 13.3 we will discuss the space MM := {singularities in one д-homotopy class}/right equivalence A2.24) and prove that it has a natural structure of a complex space. The group G% = Aut(#g°, h, S) acts on У>-1, respecting the pairing Kj and the actions of z and Э. It acts on DPmhs and DBL properly discontinuously (Theorem 10.34). The quotient DBL/G% is the moduli space for Brieskorn lattices, or more precisely, for tuples (Я|°, h, S, V>-l, Hq) up to isomorphism. One obtains a global period map M, -* DBL/GZ. A2.25) In [He2] the following global Torelli type conjecture was formulated. Conjecture 12.7 The period map fi-homotopy class of singularities. is injective for any It was proved in [Hel][He2] for all unimodal and most of the bimodal singu- singularities. The only exceptions were some subseries of the 8 bimodal series where the period map could not be determined precisely enough to see whether it is injective. In [He3] it was proved for the semiquasihomogeneous singularities with weights (i, 5,5,3) (with dim 5K = 5) andfor аЁ semiquasihomogeneous singularities with weights (j-,..., ^-) such that gcd (at, aj) = 1 for all i ф j (with dim 5K arbitrarily high). At to the present time no counterexamples are known. In section 13.3 we will see that the space ЛЛц, is locally the quotient of a д-constant stratum by a finite group. The period map Ф : 5M -*¦ DBl is a local lift of the global period map. M. Saito did not have the classifying space DBL,
'ill '¦О 226 jx-constant stratum but nevertheless he considered essentially the period map Ф, as a map into a manifold which is a classifying space for subspaces of V>~[/V"~l with the correct spectral numbers. In [SM3,2.10] he proved that itis injective if S^ is smooth. In [SM4, Theorem 3.3] he used this to show that for any 5M it is finite-to-one if one chooses 5М sufficiently small. He did not use the construction in section 11.1. With that and with the precise knowledge of coordinates on Dbl we obtain the following infinitesimal Torelli type result. Theorem 12.8 The period map Ф : 5M -*¦ Dbl is an embedding o/5M with its reduced complex structure into Dbl if SM is chosen small enough. Proof: Choose an opposite filtration U, for the singularity / and elementary sections si € Ca' as in section 11.1. We claim that any lattice К € DBl in a neighbourhood of the Brieskorn lattice Щ' e Dbl has a unique element of the type vf = A2.26) for / = 1,..., ix. One constructs it in two steps, first the principal part, then the rest. The filtration U, is also opposite to the Hodge filtration prpMHs(K) € Dpmhs- So one obtains spaces GaK as in A1.5). They satisfy the analoga of A1.6)—A1.10) and also a г A2.27) p>0 and vice versa. There is a unique elementary section sf € G^ with sf — Si € = Si + p: P>\.aj-P=at A2.28) The element vf is built from the sections sf by the same procedure as in A1.24) and the proof of Theorem 7.16. Uniqueness is clear from the construction. The sections vf generate К as а С{{Зг}}- module and as a C{z}-module. The lattice К and the coefficients у>р{К) determine one another uniquely, i The period map Dbl sends t e 5M to the lattice К е DBl p p with y[F\K) = c\f(t). Now the injectivity of the period map follows simply from the facts, that the coefficients c\- for у = 1,...,/x are flat coordinates on M (Theorem 11.9 (b)) and that 5M С {r | c^J = 0foray - 1 - ax < 0} (Theorem 12.2). 12.2 Period map and infinitesimal Torelli 227 To see that the period map is an embedding is much more difficult. We will show the existence of an index set / С {(/, ;, pY\ otj - p - a, > 0, p > 0} and of a submanifold M С M with 5M_C M such that the coefficients \v! di M d h h ffii {jp | } a c\v! nd of a submanifold С M_ |A, j, 1) € /} serve as coordinates on M and that the coefficients {c\ |A, j, 1) } yjp (i, j, p) € /} serve locally as coordinates on Dbl- That is obviously sufficient. The existence of an index set / such that {y-f | (i, j, p) e 1} serve locally as coordinates on DBl follows from the construction in [He4]. It decomposes into the index sets h - {(«. h P) e 7 I aj ~ P ~ a'" = °J and h = l ~h- The set h corresponds to coordinates for the fibres of prpMHS '¦ Dbl -*¦ DPMhs- It can be chosen explicitly as h = {(i, j, 1) I' < Д + 1 - aJ ~ l ~ «'¦ °Ь A2.29) compare [He4, Theorem 5.6] and A1.51), A1.52), A1.55), and Remark 11.10 (b). It especially contains all triples (I, j, 1) with aj — 1 — a,- > 0. The set I\ corresponds to coordinates for the base DPmhs- Unfortunately, an explicit choice is difficult and not provided in [He4]. We now have to construe t it. Let us first forget about the si and yfp and find a basis af of Я°° and coefficients f}^ with ay — p - a, = 0, which are better suited. 021 h i b lp p'4 and Because of the proof of Lemma 10.21 there exist subspaces l with all the properties in Lemma 10.18 except A0.51)+A0.52) and which give the opposite filtration Um by the analogon of A0.66). One can choose bases of all the primitive subspaces (Iq'4)x such that they together with their images under powers of N form a basis a\,..., a° with the following properties: set a°+1 := 0; there exists a map v : {1,..., ll} ->¦ {1,..., /x, д + 1} and an involution к : {1,..., /л] -*¦ {1,..., д} with N af = a%y a? € П A2.30) A2.31) A2.32) here FJ denotes the Hodge filtration for /. The basis a\,..., a° is a basis of Jordan blocks which are dual or selfdual with respect to the polarizing form S and which fit to the splittings and strict morphisms in Lemma 10.18. Note that N is an infinitesimal isometry of S. For any Hodge filtration F' € DPMHS sufficiently close to Fo*, there exists a unique basis a\,..., a^ of #°° with a,- a. eF[n-a']n U[n-ai] a? € U[n-a,]-\- A2.33) A2.34)
228 It has the form fi-constant stratum : p>0,ocj—p—ai=0 A2.35) One checks as in the proof of Lemma 10.21 that the a\,..., aM satisfy the same relations A2.30) and A2.31) with respect to S and N as the a°,..., a°. This gives relations between the coefficients p\p. The relations from N say: (a) each coefficient f}\f with a? e ImN is zero or equal to a coefficient (iff with а^ <? ImN, OS) such a coefficient #[f> with a? ? ImN is zero if a°k e kerN and a° ? кегЛР1 for some m. One eliminates with (or) all coefficients fi\f with a, e ImN. With respect to S, observe that N is an infinitesimal isometry. One checks easily that the relations from S satisfy: (y) All the nonvanishing different relations have one or two coefficients as linear terms and otherwise quadratic terms; in one relation the upper indices or sums of two upper indices are all equal; any coefficient turns up as linear term in at most one relation. In Definition 10.16 of a polarized mixed Hodge structure, the conditions (i) and (iv) (/3) are open conditions. The others are satisfied for a filtration which is defined by a basis ab ..., aM as in A2.35) whose coefficients p\f satisfy the relations from the above N and S. In view of (а)-(к), certain of the coefficients $p form local coordinates for DPMHS. We are especially interested in the behaviour of the coefficients /?{'•': (S) certain of them might be zero because of (/3), (e) some others might be zero or linearly related only to one another because of(K), (?) all others can be chosen as coordinates. Now we return to the coefficients у^\ The relations in (S) and (e) induce linear relations between the у.Ф for or, - 1 - a, = 0. The index set /, can be chosen to contain the indices of all y,y with a,- — 1 — ct\ — 0 except those which can be eliminated by these relations. The corresponding linear relations between the c\l? with Uj — 1 — a\ =0 define the submanifold M С [t e M \ c^J = 0 for a, - 1 -ai < 0}. Obviously SuCM. П 12.2 Period map and infinitesimal Torelli 229 Remarks 12.9 (a) It is not clear whether the period map Ф : 5M -> DBL is an embedding of 5M with its canonical complex structure into Dbl- The reason is that it is not clear whether SM with its canonical complex structure is embedded in the flat submanifold M which was constructed in the proof. One does not know whether the corresponding relations between the c^j with a, — 1 - a, =0 from (E) and (e) in the proof are contained in the ideal in Theorem 12.2. (b) It might be that such extra relations do not turn up or that they are always contained in the ideal in Theorem 12.2. It might also be possible to define in a natural way an a priori bigger ideal which takes such relations into account. But one would need more elaborate choices and independence of choices would be a much bigger problem. An infinitesimal refinement of Vl(O/Jf) which takes into account relations from the Jordan blocks and the weight filtration (i.e. type (/3) and (S) in the proof) has been discussed in [ChV]. (c) The sections vf in A2.26) are very similar to the sections t>,- in chapter 11. One has to replace M by DBL. The sections vf glue to generators of a free CPi x Dbl -module over P1 x DBt, which is the sheaf of sections of aflat vector bundle over C* x DBL. The restrictions to (Д x [K], @, K)) generate K. Along {oo} xDbl the sheaf has a logarithmic pole, and the restrictions to {oo} x Dbl of the sections vf are flat with respect to the residual connection along [oo}xDBL.
J ! Chapter 13 Moduli spaces for singularities The choice (?/., yO in Theorem 11.1 induces isomorphisms between spaces which are not canonically isomorphic. Section 13.1 fixes this and the compatibil- compatibilities of these isomorphisms with respect to coordinate changes and /л-constant deformations. This is used in section 13.2 for a discussion of symmetries of singularities and in section 13.3 for the (up to now) best application of the construction of Frobenius manifolds in singularity theory: the existence of global moduli spaces for singularities. This really comes from an interplay between polarized mixed Hodge structures and Frobenius manifolds, with the construction in section 11.1 as the link. 13.1 Compatibilities A choice (?/., y{) as in Theorem 11.1 for a singularity / does not only in- induce the metric of a Frobenius manifold. It also induces isomorphisms be- between several /u-dimensional spaces which are not canonically isomorphic (Lemma 13.3). These spaces and isomorphisms behave naturally with respect to /Lt-constant families (Lemma 13.4) and with respect to isomorphisms of singu- singularities (Lemma 13.6). A part of this is elementary, a part is a direct consequence of the construction in section 11.1. АЁ of it fits to the expectations. But to use it in sections 13.2 and 13.3 we need precise notations and statements. First, in Theorem 13.1 and Theorem 13.2 some results from part 1 are summarized which will be needed in sections 13.2 and 13.3. Theorem 13.1 Let {{M, t), o, e) be the germ of a massive F-manifold with multiplication о and unit field e. The group Aut((M, t), o, e) of automorphisms of the germ of the F-manifold is finite. If E is an Euler field of the F-manifold (i.e. Lie?(o) = 1 • o) such that the endomorphism TtM -» TtM, Ik?oI, is nilpotent, then Aut((M, t), о, е) — Aut((M, t), о, е, Е). 230 13.1 Compatibilities 231 Proof. The finiteness is Theorem 4.14. An Euler field is mapped to an Euler field by an automorphism of an F-manifold. If the germ (M, t) has an Euler field, then there is a unique Euler field whose action on T,M is nilpotent. This follows from Theorem 3.3. Any automorphism of the germ maps this Euler field to itself. D Theorem 13.2 Let f(x0, ...,xn) and f(x0,..., xn) be two isolated hyper- surface singularities with semiuniversal unfoldings F and F and germsofbase spaces (M, 0) and (M, 0) as in section 10.1. The germs (M, 0) and (M, 0) are germs of massive F-manifolds. (a) A germ of a biholomorphic map <p : (Cn+1, 0) -*¦ (C+1,0) (i.e. a co- coordinate change) with f — f о <р can be extended to an isomorphism of the unfoldings, that is, a pair of isomorphisms (Ф, q>M) such that the diagram commutes (C+1 x M, 0) [pru (M,0) (C+1 хМ,0) A3.1) 4>u (M, 0), and Ф|(С"+1х@},0) — ' F = F оФ A3.2) A3.3) hold. The map <pM is an isomorphism of germs of F-manifolds. The map Ф is not unique, but <рм is uniquely determined byjp. (b) For any isomorphism <pM ¦ (M, 0) -» (M, 0) of the germs of F-manifolds there exists а Ф such that the pair (Ф, ipM) is an isomorphism of the unfoldings, that is, it satisfies A3.1) and A3.3) (Ф is not at all unique). Proof, (a) The existence of an isomorphism (Ф, срм) of the unfoldings which extends cp is classic, it is essentially part of the definition of a semiuniversal unfolding (e.g. [Was], [AGV1, ch. 8]). The base map <pM is an isomorphism of germs of F-manifolds, because the Kodaira-Spencer maps, which define the multiplications, are compatible with the pair (Ф, срм) (E.10) and Remark 5.2 (v)). It is unique because the tangent map at 0 is unique and the germs of F-manifolds are so rigid by the above Theorem 13.1 (Theorem 5.4). (b) Theorem 5.6 (b)). ?
232 Moduli spaces for singularities Now we turn to the construction in section 11.1. Let f(xo,.. ¦, х„) with n > 1 be an isolated hypersurface singularity as in chapters 10 and 11 and F : (C"+1 x M, 0) -» (C, 0) a semiuniversal unfolding, H" the cohomology bundle over A x M — V (for a representative of F), 7^@) the extension from Lemma 10.2 to A x M of its sheaf of holomorphic sections, Щ' the Brieskorn lattice, H°° the space of global flat multivalued sections on Я"|(Л х {0}, 0), and F* Steenbrink's Hodge filtration on it. Consider the four д-dimensional spaces TqM, O/J/, Qj, and H°°. Here O/J/ is the Jacobi algebra and (cf. A0.87)) Й, = л The reduced Kodaira-Spencer map ao : T0M -*¦ O/Jf (cf. A0.1)) is a canon- canonical isomorphism. The space uf is a free 0/-//-module of rank 1; generators are the classes in Й/ of volume forms. One has to choose such a class of a volume form to obtain an isomorphism O/J/ = ?1/. There are no obvious natural isomorphisms between H°° and the other three spaces. Lemma 13.3 A choice (U,, y\) as in Theorem 11.1 induces the isomorphisms ., у\\ in the following diagram TcsM O/Jf 4 A3.4) The map i/f| comes from the splitting Я00 = 0p Fp П Up, the map 1/^4 is the O/Jf-module isomorphism with V4(l) = [v?] (cf A1.16) and A1.17)), the map 1/^3 is the composition of the projection Щ -*¦ ?1/ with an embedding A3.5) which will be explained in the proof. Denote by ToM A3.6) the induced isomorphism. The maps ijf\, 1/^2, and 1/^3 depend only on U., the maps iff4 and 1/^5 depend also linearly on y\, that means, for с 6 С* one has A3.7) Proof. For the definition of i/r2[U.] one has to be well aware of the construction in section 11.1 up to A1.17). The map 1/^2 is composed of the isomorphisms 1/r f1 Gr"pH° UPH°°, A1.5), A1.13), A1.4), and the restriction of 13.1 Compatibilities 233 global sections in ft@) to sections in the Brieskorn lattice Щ. Properties of fa will be discussed in the proof of Lemma 13.4 The dependence on yi of the isomorphisms follows from i/r3(yi) = [u°]. ? Lemma 13.4 Fix a choice (U., y\) as in Theorem 11.1. Consider two points h and h in the ^-constant stratum 5M С М. The spaces Я00 of the corresponding singularities are canonically isomorphic and will be identified (cf. Remark 12.3). By Theorem 11.1, the opposite filtration U. (without y\) induces aflat struc- structure on M. This yields an isomorphism <y[U.,tut2\:ThM^ThM. A3.8) One has for the singularities for t\ and ti isomorphisms as in A3.6). Denote them by ^sKU., y\), t\] and ^sKU., yi), tj\ They are compatible with A3.8): A3.9) a[U., tu ь] о fs[{U., yx), ty] = Ши., П), hi Proof. One has to review the definition of 1/Г5. The isomorphism 1/^ ' and the spaces G" of elementary sections in A1.5) depend on the Hodge filtration, which may vary for the singularities in 5M. But the composition of i/лj~ ',A1.5), and A1.13) is independent of the Hodge filtration. With A1.4) one arrives at the same isomorphism • vi A3.10) forallsingularitiesinS^.Onenowhastoseethatforanyf 6 S^ the isomorphism фС- v,- -> T,M, Vi^Si\T,M > A3.11) coincides with the composition of the restriction of u,- to the Brieskorn lattice HqQ) and the isomorphisms 1/r^1 and ap1 for the singularity corresponding to t e Sp. This follows with Lemma 11.5 (b). , D Remarks 13.5 (a) The isomorphism i/^i depends on the Hodge filtration of a singularity. Therefore, in general, one does not have a compatibility of the iso- isomorphisms 1/Г5 о i/^! [tj] : Я°° -*¦ Tu M, i = 1, 2, for two parameters t\, ti € SM with A3.8). (b) In the diagram A3.4) of isomorphisms the semiuniversal unfolding is needed only for the Kodaira-Spencer isomorphism ao. In the definition of ifo one can replace A1.13) and A1.4) by the corresponding restrictions to the slice P' x {0} С Р' x M, or one can define the section vf directly as the unique
234 Moduli spaces for singularities 1 iic x\ ¦< tl га section in Щ with principal part s,- and higher elementary parts as in formula A1.24). Lemma 13.6 As well as f consider a second isolated hypersurface singularity f = /(*o. ••-,*„) which is right equivalent to f. Choose a semiuniversal unfolding F with base (M, 0)for f and write all the other associated objects for f with a tilde. Fix a coordinate change cp : (C"+1, 0) -*¦ (C+1, 0) with f = f°<P- (a) It induces isomorphisms between the corresponding objects for f and f, which respect all the canonical additional structures, <pM : ((M, 0), o, e, E) -» ((M, 0), o, 7, E), 4>Jac : (O/Jf, mult., [/]) -* (O/Jj, mult., [/]), cpcro : (П/, Jf) -* (П?> Jy), Van : {Щ, z, d~\ Kt) -* (Щ, z, Ь~\ Kj), <pcoh : (й|°, h, S, F') -* (H™, h, S, F'), (cf. section 10.6 for the residue pairing Jf, the polarizing form S and the form Kf). (b) Choose (?/., yi) as in Theorem 11.1 and set U. := <pcoh(U.) in H°°, Y\ := (Gr^cpcohXyO. The diagram A3.4) for (/, ([/., yx)) and the diagram A3.4) for (/, (U., y\)) are compatible with the above isomorphisms and with the induced isomorphisms. The same holds for the embeddings fo and \jf2 from A3.5). The map <рм is an isomorphism of Frobenius manifolds with respect to the metrics on M and M induced by(U.,y\) and ([/., y\). Proof. This holds, because all the considered objects and structures and also the construction in section 11.1 are essentially coordinate independent. Theorem 13.2 provides the uniqueness of cpM. For the other isomorphisms one does not even need to consider the semiuniversal unfoldings because of Remark 13.5 (b). A nontrivial point is to only formulate one or several nat- natural definitions for the isomorphisms <рм, ••¦', <pCoh and to check that all are compatible. See Remark 13.7. ? Remark 13.7 For example, <рв„ can be defined by pulling back differential forms with (ср~1)*. It can also be recovered from <рс„&: By sections 7.1 and 10.6, the tuple (H°°, h, S) determines the structure (V>, z, d~\ Kf). The 13.2 Symmetries of singularities 235 isomorphism cpcoh induces an isomorphism from (V> ', z, dz ', Kf) to the cor- corresponding structure for /. It maps Щ to H^. Another example: 0/7/ is naturally embedded in End^/_ adjoint(Q,f, Jf) by the action on ?2/. Knowing this embedding for / and /, one can recover cpJac from cpGm. 13.2 Symmetries of singularities The following is motivated by the results of Slodowy [SI] and Wall [Wall] and an extension of them. Remarks on the relations will be made below. Let / : (C+1, 0) -» (C, 0) be an isolated hypersurface singularity with n > 1, П := {<p : (C"+1,0) -> (C+1,0)biholomorphic} the group of all coordinate changes, and TZf = [tp e U \ f о q> — f) the group of symmetries of/. The group of symmetries acts on all the objects associated to the singularity. The interrelations presented in section 13.1 result from heavy machinery and profound facts, but now they make it easy to analyze the actions of the symmetry group. The group TZf is oo-dimensional, but the group of fc-jets jkR-f is an algebraic group for any к > 1. Let A3.12) be the finite group of components of j\R.}. The following is classical. Lemma 13.8 The kernel kev(jkTZf -> jiTZf) is unipotent. The groups jkTZf and j\VJ have the same number of components. Proof. The group jkTlf acts faithfully on 0c«+i,o/m*+\ where m is the maxi- maximal ideal. An element of the kernel acts trivially on m/m2 and thus on m' /m'+1 for any / < k. It is unipotent. The kernel is unipotent and thus connected. ? We use the following abbreviations for the groups of automorphisms of different objects associated to the singularity. AutM = Aut((M, 0), о, е, Е), A3.13) Autyac = Aut(C/7/, mult., [/]) = AutGbM, o, ?|0), A3.14) Gz = Aut(H^,h,S), A3.15) StabGz(F") = Aut(H^, h, S, F'), A3.16) StabCz(flSO = Aut(ff|°, h, S, V"-1, Hg). A3.17)
236 Moduli spaces for singularities Here (M, 0) is the base of a semiuniversal unfolding, a germ of an F-manifold with Euler field, O/Jf is the Jacobi algebra, Я|° the lattice in the space Я°° of global flat multivalued sections in the cohomology bundle over a punctured disc, h its monodromy, S the polarizing form from section 10.6, Щ the Brieskorn lattice, and F* Steenbrink's Hodge filtration on H°°. One can see that Gz is isomorphic to the automorphism group of the Milnor lattice with Seifertform[He4]. The group Gz acts on(V>~l,z, d~l, Kf)(com- pare sections 7.1 and 10.6). It acts properly discontinuously on the classifying spaces DPmhs and DBl (Theorems 10.22, 10.34). Therefore the groups Stabcz(#o') С Stabcz(F*) are finite. The group Autjv/ is also finite (Theorem 13.1). Because of Theorem 13.2 and Lemma 13.6 there are canonical homomorphisms from the group Kf of symmetries of / to the groups in A3.13)—A3.17). Denote deto : Uf -»¦ C*, <pv+ det (—\ @). A3.18) Because of the splitting lemma one can transform any singularity by a coordinate change to the form in Theorem 13.9. Theorem 13.9 Suppose that f = g(x0, ...,хт)+х„+1-\ Yx\withg e m3 for some m < n. (a) The canonical homomorphismsfromiV to the groups in A3.13)—A3.17) factor through Rf. (b) If n = m then the map Rf -*¦ Autji/ is an isomorphism. Ifn>m then Rf = Rgy. %2, the map Rf —> Aut*f is two-to-one, and the kernel is generated by (x0,..., х„_ i, х„) h* (xo,..., xn-1, -xn). In both cases (n > m), the map Rf —> Autyac is induced by the embedding of Aut« in AvA{TqM, o, ?|o) = AutJac. (c) Ifn = m then Rf = j{Rt and any maximal reductive subgroup of' j^RJ is isomorphic to it. Ifn > m then any maximal reductive subgroup of j]Ji.f is isomorphic to Rg x O(n — m, C). (d) The map deto factors through Rf. The action of Rf on H°° is obtained from the action on TqM by twist with det^. More precisely, there is an isomor- isomorphism f6 : H°° -» T0M such that for all <pe1lf <Pcoh = V*6 ° det0 («0 A3.19) Here <pcoh € Gz and <рм ? Aut^ are the induced isomorphisms. (e) The map Rf —>• Stabcz(HQ) is injective, so Rf acts faithfully on the Milnor lattice. 13.2 Symmetries of singularities 237 Proof, (a) The group Aut^ is embedded in Aut(T0M, o, ?|0) = Autyac, be- because it is finite. The actions of IZ? on ToM and on O/Jf are isomorphic via the Kodaira-Spencer isomorphism ao. The action on O/J/ depends only on sufficiently large fc-jets and depends continuously on them, the action on TqM is finite. Therefore both factor through Rf. The same argument works for H°° and fi/ with a special choice of U,: The opposite filtration U^ from Deligne's /?•? on H°° (cf. A0.66)) [/(°>= A3.20) is invariant under <pCOh for any <p e Л?, because the 1рл are invariant by their definition (cf. A0.50)). It induces an isomorphism ^3 о ф{: Н°° —»• fi/ by Lemma 13.3. The actions of Tlf on Й/ and on H°° are compatible with it by Lemma 13.6. But StabGz(//o') is finite and the action on fi/ depends continu- continuously on sufficiently high &-jets. So both factor through Rf. (b) and (c) First, suppose n=m. Then /em3 and there is a natural surjec- tive map m/7/ —>• m/m2. The group action of j\Rf on m/m2 is faithful. It factors through Rf by (a). Thus j\Uf ~ Rf and Rf acts faithfully on O/Jf. Therefore Rf -» Aut« is injective. It is surjective because of Theorem 13.2 (b). Because the kernel ksr(jkTZf -> jiTZf) is unipotent (Lemma 13.8), any reductive subgroup of jkTZf is mapped injectively to Rf. Now, suppose n > m. A symmetry <p e 72/ acts on Cc»+',o by h ну h о <р~х. It leaves the Jacobian ideal invariant. The ideal m2 + 7/ = m2+(xm+\,... ,xn) is invariant under the action of 72/. There is a natural homomorphism ]{ Aut m + (xm+ x Aut I ...,xn)J \ m2 A3.21) The kernel is unipotent, the image is isomorphic to Rg x O(n — m, C), because j2f = x^+l + ¦ ¦ ¦ + *2. Therefore Rf = Rg x Z2. The other statements are also now clear. (d) Choose any generator y\ of the 1-dimensional space Gr^_a|]W°°WB|. By Lemma 13.3 (U®\ y\) induces a class of a volume form [d°] e Я/ and an isomorphism 1/4 : O/Jf -» fi/, [1] н* [v°], of C/7/-modules. Also fiiyi) = [«?]• The actions on Gr17 Я°° and fi/ of a symmetry <p e 72.-^ are compatible with 1Д3 and have yi and [d°] as eigenvectors. The eigenvalue is det^'(#>), because [v\\ is represented by a volume form. By Lemma 13.6 the actions <pjac and <paf are compatible with 1^4 with the twist det^' (<p\ ° <PJac = <Pa, A3.22)
238 Moduli spaces for singularities Consider A3.4) and define := a ' ' о fa о fa. The compatibilities of the actions of <p with a0 and fa о fa show that the twist extends to A3.19). The map deto factors automatically through Rf. (e) The kernel ker(/J/ -> Gz) is contained in the kernel of deto because of the action on С ¦ yx in (d). Then by A3.19) it is the intersection of the kernels of deto and Rf -> h.\AM. But this is trivial. D Remarks 13.10 (a) Slodowy [SI] considered a real singularity with a given compact group of symmetries and showed that one can construct a semiuniversal unfolding on which the group acts. He also showed that such a group is finite if /em3. Compact corresponds to reductive in the complex case. So his results are close to Theorem 13.9 (b) and (c). But he did not have the uniqueness of the action on the basis (M, 0), which follows from the rigidity of massive F-manifolds (Theorem 13.1), and he did not consider Rf and AaX.M- (b) Part (d) in Theorem 13.9 is due to Wall [Wall], also the way in which this is proved. But he did not have the up to a scalar canonical isomorphism fa : H°° -> TqM, which comes from the construction in section 11.1 and from Deligne's /"•?. (c) In [SI] and [Wall] one starts with a given (compact or) finite subgroup of TZf. The group Rf and its liftability to a subgroup of lZf are not considered. But this liftability can be shown easily, following arguments in [WaI2] (see also [Muel]): It is known that the group Rf can be lifted to a subgroup Rf in the algebraic group jkTZf for any k. Let к > /м +1. Within the fc-jets one carries out the usual averaging procedure to find coordinates on which the group Rf acts linearly and obtains а к -jet jkip of a coordinate change such that jkcp~l о Rf о jk<p acts linearly on O/mk+l and respects jkf о jkip. The last group lifts to a group of linear coordinate changes and respects jkf о jk<p, considered as a polynomial. This polynomial is right equivalent to /, because / is д + 1-determined. In the case of a quasihomogeneous singularity / one can calculate the group Rf using the following characterization. Theorem 13.11 Let f e C[x0,..., xn] be a quasihomogeneous isolated sin- singularity with weights wo,...,wn 6 @, |] П Q arid degree 1. Suppose that wo < • • • < wn_] < I (then /em3 if and only if wn < \). Let Gw be the algebraic group of quasihomogeneous coordinate changes, that means, those which respect C[*o, • • •, xn] and the grading by the weights w0,..., wn on it. Then Rf = StiibGJf). A3.23) 13.2 Symmetries of singularities 239 Proof. The group StabG.,(/)is nnite ЬУ [GrHP, Proposition 2.7] (the proof is similar to that in [SI, D.6)]). In [GrHP, Theorem 2.1] it is proved that any symmetry cp e TZf has weighted degree > 0, that means, the г-th component cpi e О does not contain monomials of weighted degree < w,. The degree 0 part of any symmetry is an element of StabG,,,(/). One can rewrite Lemma 13.8 for weighted jets and for StabGl,.(/) instead of }{RJ. But the groups of jets and of weighted jets of symmetries are con- contained in one another for suitable high degrees. They have the same number of components. This shows A3.23). D Conjecture 13.12 The map Rf -у StabGz(#o) is an isomorphism for any iso- isolated hypersurface singularity f(x0, .. ¦, xn) with n > 1 and multiplicity 2. Theorem 13.13 Suppose that the map Rf ->¦ StabGz(Wo') is an isomorphism for one singularity f(x0, ...,xn) with n > 1 and multiplicity 2. Then the corresponding map is an isomorphism for any singularity in the (sufficiently small) ix-constant stratum 5M of f. Proof. We may choose a (sufficiently small) representative F of the semiuni- semiuniversal unfolding, a representative in Hf for each element of Rf, and extensions of these representatives to automorphisms of the unfolding F in the sense of Theorem 13.2 (a). Consider the period map S^ -> DBL, t м>- Н?Ц), from section 12.2. If SM is sufficiently small, then StabGz(#?(O) С StabGz(Ho'@))foranyr е S^ because Gz acts properly discontinuously on DBL (Theorem 10.34). By assumption, an element cpcoh 6 StabGz(Wo (О) С StabGz(tfov(O)) = Rf is induced by an element of TiJ. It induces an automorphism cpM of M. By the compatibility of the morphisms in Lemma 13.6, <pco/,(ffo'@) = Щ(<Рм^У). But the period map is injective (Theorem 12.8). Therefore <pM(t) = t. The automorphism of the unfolding restricts to a symmetry of the singular- singularity F,. This symmetry induces cpcoh. Therefore RFl -у StabGz(Wo'@) is an isomorphism. ^ Remarks 13.14 (a) Conjecture 13.12 complements the global Torelli type conjecture 12.7. Together with the infinitesimal Torelli type theorem 12.8 and Theoreml3.15belowitwouldsaythatlocallytheperiodmapyHM -*¦ DBL/GZ is an embedding. The only obstructions to Conjecture 12.7 would be possible intersections of the images of disjoint pieces of Мц. (b) The conjecture is true for the weighted homogeneous singularities with weights (^,..., ^) where gcd(ai,aj) = 1 for i ф j.
X I "О !-< л/У О от 240 Moduli spaces for singularities For these singularities all eigenspaces of the monodromy are 1-dimensional and the orders of their eigenvalues can be put into certain chains with one biggest order. This is used in [He3, Proposition 6.3] to show that StabcT(#g') = G% = {±hk | к e Z). This is isomorphic to StabCi/,(/) if an = 2. (c) The same reasoning as in (b) applies to the Ад., D2k+\, Ek, and to 22 of the 28 quasihomogeneous unimodal and bimodal exceptional singularities, but not to D2k, Zn, Qii, Ul2, Zuj, 2,6, Ul6. Starting with a Coxeter-Dynkin diagram, I checked Conjecture 13.12 for Q\2, which has some 2-dimensional eigenspaces and Stabc,,(#o') = ^z = {±hk | A e Z} x Z2 (= Rf for n > 3, = StabC]i,(/) for и = 3). 13.3 Global moduli spaces for singularities In this section we present our best application of the construction of Frobenius manifolds for hypersurface singularities: Theorem 13.15 gives a global moduli space for the singularities in one /x-homotopy class. The three main ingredients in the proof are the F-manifold structure on a base M of a semiuniversal unfolding of a singularity, the construction of fiat structures on M by choice of opposite filtrations, and the period map 5Д —> Dpmhs from a ц.-constant stratum to the classifying space for polarized mixed Hodge structures. So it grows from an interplay between Frobenius manifolds and polarized mixed Hodge structures with the GauB-Manin connection as the intermediary. The construction of moduli spaces for singularities starts with Mather's the- theory of jets of singularities ([Mathl][Math2], cf. [BrL]). The Ar-jet of a function f e О = 0<c»+i,o 1S tne Class jkf € O/mk+1; here m С О is the maximal ideal. The action of the group П = {<p : (C"+t, 0) ->¦ (O+I, 0) biholomorphic } of coordinate changes on m2 С О pulls down to an action of the algebraic group jkTZ of A:-jets of coordinate changes on m2/ini+1. An isolated hypersurface singularity /em2 with Milnor number /x is /x +1- determined, that means, any function g with yM+, g = j^+if is right equivalent to /. Therefore for A > /x + 1, the set of all 7?.-orbits in m2 of singularities with Milnor number /i is in one-to-one correspondence with the set of /^-orbits of their A-jets in m2/m*r+l. Fix/i and A: > /x + l.Thecodimensionof the orbit j1lK'/i/mm!/ni'+' fora singularity/with/x(/) = /x is д — l. The union of all orbits with codimension > ijl — 1 is an algebraic subvariety in m2/m*+l. The set {jkf | /x(/) = /x) is Zariski open in it and thus a quasiaffine variety. It decomposes into algebraically irreducible components and into (possibly bigger) topological components. 13.3 Global moduli spaces for singularities 241 Each topological component corresponds to a /x-homotopy class of singular- singularities. The singularities / and g in m2 with /x(f) = fi(g) = /x are /x-homotopic, f ~м g> if and only ^tnere is a family / 6 m2, t e [0, 1], with jx{f) = /x, /o = /, /, = g, such that the coefficients of the power series f, depend contin- continuously (or, equivalently, C00 or even piecewise real analytic) on t. A singularity / with /x(/) = /X is /i-homotopic to its A-jet by the proof of the finite determi- nacy, and A-jets in one topological component of the set above (Л / | /x(/) = /i} are obviously д-homotopic. Finally, the group j^TZ is connected and acts on each topological component. For a singularity /6 m2 with /x(f) = /x, denote by C(A, /) the topological component of [jkf | /x(f) = /x] which contains /. We summarize: the map {g e m2 | g ~M f)/n -> C(A, f)/jkU A3.24) is bijective, the algebraic group jkTZ acts on the quasiaffine variety C(A, /) С m2/m*+1. Theorem 13.15 Fix /x, к > /x + 1 and a singularity / e m2 С Сс»+' ,o with (a) The quotient М„. of the map ж : С (А, /) -»• C(k, f)/jkU =: 74M is an analytic geometric quotient. That means, the quotient topology is Hausdorff and the sheaf A3.25) A3.26) induces a reduced complex structure on (b)Thegermat[jkf]is ц, [jkf]) = OS*, 0)/Aut((M, 0), о, е, Е). Here ((M, 0), о, е, Е) is the base space of a semiuniversal unfolding of f with its structure as a germ of an F-manifold (with Euler field), (SM, 0) С (М, 0) is the ^.-constant stratum. (c) The canonical complex structure on the ix-constant strata from Theorem 12.4 induces a (possibly nonreduced) canonical complex structure on M^. The proof will be given after Remark 13.19. Results of Gabrielov [Ga] and Teissier [Tel] will allow semiuniversal unfoldings to be considered instead of spaces of A:-jets. Then Theorem 13.17 shows that the quotient topology is Hausdorff. Theorem 13.18 gives the reduced complex structure and A3.26). Remark 13.16 (a) All orbits in C(k, /) have the same dimension and are closed. But this by no means implies that the quotient is Hausdorff. For example,
a 242 Moduli spaces for singularities the quotient of C2 - {0} with the action of C* ^ GL{2, C), x i->- (J^,),on it is not Hausdorff, the punctured coordinate planes cannot be separated. (b) One may ask whether M^ is even an algebraic geometric quotient. But it is not at all clear how one could approach this question. In general, the group jkTZ is not reductive and the variety C(k, /) not affine. (c) I expect that M.^ with the canonical complex structure in Theorem 13.15 (c) is a coarse moduli space for an appropriate notion of д-constant deformations over arbitrary bases. But this has still to be worked out and checked. Theorem 13.17 Let f and f e 6>c«+> ,o be two isolated hypersurface singulari- singularities with (representatives of) base spaces M and M of semiuniversal unfoldings and with ^-constant strata S^C M and 5M С М. Suppose that there are two sequences(tj)iej<iand(tj)jew with Ъ e S^.Ti e 5M, f,-—»¦ OandTj —»¦ Ofori —> oo, such that for each i € N the singularities which correspond to ti and ti are right equivalent. Then f and f are right equivalent. Proof. We may suppose n > 1, because for и =0 there are only the Ад- singularities. Let F and F be (representatives of) semiuniversal unfoldings of / and /over M and M. Denote by F,,. = F|(C"+1 x {f,}, (x(i), f,-)) the singularity which corresponds to tt; here x(l) is its singular point and F,,((jt(l), r,)) = 0 by definition of 5M. Define Fj. analogously and choose a coordinate change (pi : (Cn+\xm)-+ (Cn+1,3c<l'))with Ftj = Т^ощ. There is no possibility of controlling directly the sequence (#>,),€n of coor- coordinate changes and finding a limit coordinate change for / and /. But we have induced sequences of isomorphisms on several related objects which can be controlled when they are seen together and which give the desired information. First, the germ F : (C+1 x M, (x(i), tt)) ->¦ (Cn+1,0) is a semiuniversal unfolding for Fti, the same holds for F and Fjt. By Theorem 13.2 (a) q>t induces an isomorphism of germs of F-manifolds We will show that a subsequence tends to a limit isomorphism <Роо,м ¦¦ W, 0) -> (M, 0) A3.27) A3.28) of germs of F-manifolds.With Theorem 13.2 (b) or with Scherk's result ([Sche2], cf. Remark 5.5 (iv)) one concludes that / and / are right equivalent. 13.3 Global moduli spaces for singularities 243 In order to control the sequence ((pi,M)ieN, we need the flat structures and their construction on M and M. Let H°° denote as usual the space of global flat multivalued sections in the cohomology bundle over a punctured disc for /. It is canonically isomorphic to the corresponding spaces for the singularities in the //.-constant stratum 5M and will be identified with them. It is equipped with the lattice #^°, the monodromy h, the polarizing form S (cf. section 10.6) and for each t e 5M with Steenbrink's Hodge filtration F'(t) of the corresponding singularity. The period map (cf. section 12.2) DpMHS, t н> F'(t) A3.29) to a classifying space for polarized mixed Hodge structures on H°° is holomorphic. Denote the corresponding objects for / by H°°, Ш%, h, S, and F*G) for 7 e S/j.. Choose any isomorphism X : (Я|°, A, S) -> (fl|°, h, S). One obtains a holomorphic period map ' PMHS The coordinate change <p-t induces (pitCoh : : #° A3.30) A3.31) tf~ which maps F*(f,-) to F*(^). Therefore x ° <Pi,coh e Gz = Aut(H^, h, S) acts on Dpmhs with X о <Р1,с = $PMHs(ti). A3.32) Now first note that ФPMHs(ti)-+ *pmhs@) and ^pMHsffi)^-^pmhs@) for г ->¦ oo, and second that the group Gz acts properly discontinuously on Dpmhs (Theorem 10.22). Therefore there exists an infinite subset / с N and an isomorphism cpcoh : H°° -* Я°° with and = <Pcoh for i e I <Pcoh(F'@)) - F @). A3.33) A3.34) So / and f have isomorphic polarized mixed Hodge structures. (One can apply the same arguments to DBL instead of DPMhs- If the global Torelli type conjecture 12.7 were proved one could stop here.) Choose a monodromy invariant filtration U. on H°° which is opposite to F*@) and a generator y\ of the 1-dimensional space С-г^_а|]Я°?2„,в1. By Theorem 11.1 the pair ([/., yi) induces a flat metric g on M.
1] 2 :=! a \ 244 Moduli spaces for singularities The filtration U. is opposite to F'(t) for all t 6 5M (sufficiently close to 0). The construction in section 11.1 gives the same metric g for all germs (A/, t) with t e Бц. The image (?/., у i) under ^c0/, satisfies the same properties with respect to F (t) for t e 5M and induces a flat metric g on A/. By Lemma 13.6 (b), the maps cpi<M, i e I, are isomorphisms of germs of Frobenius manifolds <Pi,M-({M,ti),o,e,E,g)-+ ((M,7i),o,e,E,g). A3.35) Even more, by combining Lemma 13.4 and Lemma 13.6 (b) one sees that they differ at most by translations (with respect to the flat structures). Thus for sufficiently large i € I, the map ц>^м extends to a neighbourhood of 0 e M, and <pi,m@) —> 0 for i -*¦ oo. If onehad^,?w@)/Oforanarbitrarilylarge/ 6 /thentheset{^,,«@) \i € I] would cluster around 0 e M. The germs ((Л/, <р,-,м@)), о, ?, E) of F-manifolds with Euler field would all be isomorphic. This is not possible by Corollary 4.16. (Also, the singularities for the parameters <р,,л/@) б 5М would all be right equivalent. With the proof below of Theorem 13.15 and with a closer look at the action of the algebraic group jklZ on the algebraic variety C(k, f) one finds that this is also not possible.) Therefore <р,,м@) = 0 for large i e I. This is the limit isomorphism from A3.28). One finishes the proof with Theorem 13.2 (b) or Scherk's result. D Theorem 13.18 Let f : (C+1, 0) -* (C, 0) be an isolated hypersurface sin- singularity. There exists a representative M of the base space (A/, 0) of a semiu- niversal unfolding of f with the following properties. The finite group hut м '¦= Aut((A/, 0), о, е, Е) of automorphisms of the germ of an F-manifold acts on M. Ift and t are in the (i-constant stratum S^CZM and (p : ((A/, t), о, е, Е) —> ((A/, t), о, е, Е) is an isomorphism of germs of F-manifolds, then <p e Aut«. Proof. We may suppose n > 1. Assume that such a representative does not exist. Then one can choose (for some representative M) two sequences (f,),?N and (f|)ieN with f,-, f,- 6 S^, f,- —>• 0 and t-, —> 0 for i —*¦ oo, and a sequence of isomorphisms ((A/, ti), o, e, E) -* ((А/,?)), о, е, Е) A3.36) with cpi g Autw for any i. By Theorem 13.2 (b), for each i the singularities which correspond to Ц and t, are right equivalent and there exists a coordinate change between them 13.3 Global moduli spaces for singularities 245 which induces щ. But then the proof of Theorem 13.17 provides an infinite subset / С N such that all щ for i e I are equal and contained in AutM. A contradiction. D Remark 13.19 This resultdoes not extend to t, t e M — 5M. For example, in the case of an A^-singularity, the Lyashko-Looijenga map (section 3.5 and E.49)) M -+ ©* is finite of degree (/x + 1)д~'. This implies that a generic germ of a semisimple F-manifold with Euler field turns up at (д + 1)M~' different points in M. But the group Aut^ only has order jjl + 1. So most of the isomorphisms of the germs of F-manifolds do not extend to 0 e M. Proof of Theorem 13.15: The jt7^-orbit of jkf in m2/m*+1 has codimension /i - 1. A transversal disc for / is an unfolding F : (C"+1 x M, 0) -* (C, 0) of / with the following properties: M С См~' is a neighbourhood of 0; for any parameter a = (au ¦ ¦ ¦, g^-i) 6 M the function germ Fa is Fa = F\(C"+l x {a}, @, a)) 6 m2; A3.37) of course Fq = /; the map Tk:M-+ m2/m*+l, а н> jkFa is an embedding and the image Tk(M) intersects jklZ ¦ jkf transversally in jkf. If we choose a smooth germ (R, id) С (JkR-,id) which is transversal to the stabilizer of jk f then the natural map of germs (R, id) x (Tk(M) П C(k, /), jkf) -* (C{k, /), jkf) A3.38) is an isomorphism. We need the germ (Tk(M) П C(k, /), jkf) to be isomorphic to the ^-constant stratum EД, 0) С (А/, 0) of /. It will follow from results of Gabrielov and Teissier. Gabrielov [Ga] constructed explicitly a semiuniversal unfolding F : (C"+l x A/, 0) -» (C, 0) of /, a transversal disc F for /, and an embedding (|, т) = (?o, ¦ • ¦, I,,, t,, ..., v> : (A/, 0) -»¦ (C+1 x M, 0) A3.39) with ). A3.40) The critical space (C, 0) С (C+1 xM, 0) (cf. section 10.1)and the intersection С П F-'@) are smooth, С = prM(C П F~40)) С М is the discriminant. Gabrielov's construction yields an isomorphism (§, r): (M, 0) ^ (СПГ'@),0) A3.41)
246 Moduli spaces for singularities 3  & б which maps T^{Tk{M) П C(Jfc, /), j*/) to the preimage in (С П F-'(O), 0) of the д-constant stratum EM, 0). The projection to the /j,-constantstratum is clearly ahomeomorphism. Teissier [Tel, §6] showed that it is an isomorphism. He used the fact that the map CnF"'@) —> V is isomorphic to the development of the discriminant (cf. for example section 3.5) to construct a section (M, 0) -» (C"+1 x M, 0) which contains the preimage ofE^, 0) in (С П F~'@), 0). Therefore the germs {T^{M) П C(&, /), j*/) and EM, 0) are isomorphic. Now Theorem 13.17 shows that the quotient topology of M^ = C(k, f)/jkR. is Hausdorff. This is the first of two conditions in a criterion of Holmann [Hoi, Satz 17] for the existence of an analytic geometric quotient. The second condition is the existence of holomorphic functions in a neighbourhood of j^f in C{k, /), which are constant on the y'tTJ.-orbits and which separate points in different orbits. By A3.38) and the above isomorphism of germs it is sufficient to show the existence of such functions on 5M. But in that case itfollows from Theorem 13.18 and from the classical result of Cartan [Ca] that a quotient by a finite group is a reduced complex space. This shows parts (a) and (b) in Theorem 13.15. The canonical complex structure on 5Д from Theorem 12.4 is invariant un- under Aut«, for example because these automorphisms respect the Frobenius manifold structure from t/^0' and the canonical complex structure is com- completely determined by this Frobenius manifold structure A st step in the proof of Theorem 12.4). One can apply locally [Каир, 49 A. 16] and glue the complex structures. This gives part (c). ? Remark 13.20 (a) Consider the local period map Ф : S^ —> Dbl for a singu- singularity /. If <p € TV is a symmetry then the action of cpM on 5M С М and the action of (pcof, on ФEМ) С Dbl are compatible. Now Theorem 13.15 shows that the global period map Мц -» DBL/Gz is a map between reduced complex varieties. (b) A part of the splitting lemma says that two singularities f(xo,..., xn) and g(xo,..., xn) are right equivalent if and only if / + x^+x and g + x%+l are right equivalent (e.g. Remark 5.7 (i)). Therefore the global moduli space М„(/) embeds into M^f + x*+l). It would be very desirable to know the answer to the following very weak form of the Zariski conjecture. Do the singularities in a д-homotopy class have either all multiplicity 2 or all multiplicity > 3? If the answer is yes, M^if) and Мц(/ + х%+1) are isomorphic, if no, the second one would be larger in some cases. 13.3 Global moduli spaces for singularities 247 (c) The answer to the question is yes for semiquasihomogeneous singularities and for curve singularities. An interesting related result has been shown by Navarro-Aznar [Nav]: sup- suppose f(x0, ...*„) and g(xo,..., х„) are singularities with the same topological type and suppose that the rank of the Hessian of /, rank( g^r)(O), is odd (this implies that its multiplicity is 2). Then also the multiplicity of g is two. This implies that two surface singularities with the same topological type both have multiplicity two or both have multiplicity > 3. (But precisely in the surface case it is not yet clear whether д-homotopy implies the same topological type, cf. [LeR]). (d) If the answer is yes for some д-homotopy class then the groups Rf for singularities / in this д-homotopy class satisfy a semicontinuity property. Then by Theorem 13.18 and Theorem 13.9 the groups RFl for the singularities in the /j,-constant stratum of a fixed singularity / are (isomorphic to) subgroups of Rj. For semiquasihomogeneous singularities the semicontinuity property of Rf has been proved previously in [Mue2].
Chapter 14 Variance of the spectral numbers Section 14.3 gives a surprising statement on the spectral numbers of quasi- homogeneous singularities. It comes from properties of the G-function of a semisimple or massive Frobenius manifold. General remarks, the definition of Dubrovin, Zhang, and Givental and some properties of it are presented in section 14.2. The socle field of a Frobenius manifold is related to the simpler part of this G-function. It is more or less known, but not treated systematically in the literature. This is provided in section 14.1. 14.1 Socle field A Frobenius manifold has another distinguished vector field besides the unit field and the Euler field. It will be discussed in this section. We call it the socle field. It is used implicitly in [Du4, Theorem 1.1] and [Gi7]. Let (M, o, e, g) be a manifold with a multiplication о on the tangent bundle, with a unit field, and with a multiplication invariant metric g. We do not need flatness and potentiality and an Euler field in the moment. Each tangent space T,M is a Frobenius algebra and splits uniquely into a direct sum of Gorenstein rings (cf. section 2.1) /@ k=\ with maximal ideals m,,* С (TtM\ and units ek such that e and thus (T,M)jo(T,M)k = g((T,M)j,(T,M)k)={0) for ]фк. 248 A4.1) A4.2) A4.3) 14.1 Socle field 249 The socle Ann(r,W)l(m,,t) is 1-dimensional and has a unique generator H,k which is normalized such that g(ek, H,,k) = dim(T,M)k. A4.4) The following lemma shows that the vectors ^t H, k glue to a holomorphic vector field, the socle field of (M, o,e,g). Lemma 14.1 For any dual bases X\,..., Xm and X\,..., Xm of T,M, that means, g(X,-, Xj) = <5,;, one has '(О Е *=i A4.5) Proof. One easily sees that the sum J2 X,- о X,- is independent of the choice of the basis X\,... ,Xm. One can suppose that l(t) = 1 and that X\,...,Xm are chosen such that they yield a splitting of the filtration T,M D mt,\ Э m^, Э Then g(e, X, о X,) = 1 and g(m,,i, X,- о X,) = g(X, о mM, X,-) = 0. ThusX,oX, = ?#,,,. D It will be useful to fix the multiplication and vary the metric. Lemma 14.2 Let (M, о, e, g)bea manifold with multiplication о on the tangent bundle, unit field e and multiplication invariant metric g.For each multiplication invariant metric g there exists a unique vector field Z such that the multiplication with it is invertible everywhere and for all vector fields X, Y g(X, Y) = g(Z о X, Y). The socle fields H and H of g and g satisfy H = ZoH. A4.6) A4.7) Proof. The situation for one Frobenius algebra is described in Lemma 2.2. It yields A4.6) immediately. Formula A4.7) follows from the comparison of A4.4) and A4.6). D Denote by X i-> H о X A4.8)
250 к t I z • -oil ; Or И ! Ir Variance of the spectral numbers the multiplication with the socle field H of (M, o, e, g) as above. The socle field is especially interesting if the multiplication is genetically semisimple, that means, generically l(t) = m. Then the caustic K. = {t e M \ l(t) < m) is the set where the multiplication is not semisimple. It is the hypersurface /C = det(#opr'@). A4.9) In an open subset of M — /C with basis e\,..., em of idempotent vector fields the socle field is l -e;. A4.10) It determines the metric g everywhere because A4.10) determines the metric at semisimple points. If (M, o, e) is an F-manifold then each germ ((M, t), o, e) decomposes uni- uniquely into a product of irreducible germs of F-manifolds (Theorem 2.11). This extends the infinitesimal decomposition A4.1). At semisimple points one has the product А"г if m 1-dimensional germs of F-manifolds. The 2-dimensional irreducible germs are classified in Theorem 4.7, those with generically semisimple multiplicationformaseries/2(n), n > 3, with /2C) = A2. Theorem 14.3 Let (M, o, e, g) be a massive F-manifold with multiplication invariant metric g. Suppose that at generic points of the caustic the germ of the F-manifold is of the type I2(n)A"'~2. Then the function det(Hop) vanishes with multiplicity n —2 along the caustic. Proof. The manifold M = O" with coordinate fields 5,- = ^- and multiplication defined by 5) о &2 = t 8\, A4.11) A4.12) A4.13) 5,- о 8j = вув, if (i, j) i {A, 2), B, 1), B, 2)} is an F-manifold with a global decomposition С2 к С x • • • x С of the type /гСяМ^2-^61111^6^ for the components are 5|, <$з> ¦ ¦ • > 5,„, the global unit field is e = 8\ + S3 -\ + ?„,, the caustic is /C = {t \ t2 — 0}. The idempotent vector fields in a simply connected subset of M — /C are 1.1 2' e, = 5,- for i > 3, e\/i = -Si ± -t2 2 82, A4.14) A4.15) 14.2 G-function of a massive Frobenius manifold 251 canonical coordinates there are 2 " mi/2 = tx ± -q, A4.16) Ui=tj for«>3. A4.17) A multiplication invariant metric g is uniquely determined by the 1-form e = g(e,.). Because of A4.7) it is sufficient to prove the claim for one metric. We choose the metric with 1-form eE,) = l-5n. A4.18) The bases 5i, 82, 53,..., 5m and 82, 8\, 53,..., 8m are dual with respect to this metric. Its socle field is by Lemma 14.1 H = 282 + 83 + ¦ ¦ ¦ + 8П: A4.19) 2 and satisfies det(#op) = If an F-manifold M has at generic points of the caustic germs of the type I2(n)A™~2 then the set of nongeneric points is empty or has codimension > 2 inM. ? 14.2 G-function of a massive Frobenius manifold Associated to any simply connected semisimple Frobenius manifold is a fas- fascinating and quite mysterious function. Dubrovin and Zhang [DuZl][DuZ2] called it the G-function and proved the most detailed results for it. But Givental [Gi7] studied it slightly earlier, and it originates in much older work. It takes the form G(t) =\ogr,-± log J A4.20) and is determined only up to addition of a constant. First we explain the sim- simpler part, log J. Let (M, о, е, Е, g) be a semisimple Frobenius manifold with canonical coordinates u\,...,um and flat coordinates t\,...,tm. Then / = i ¦ constant A4.21) is the base change matrix between flat and idempotent vector fields. One can rewrite it with the socle field. Denote щ := g(e,-, e,) and consider the basis v\,..., vm of vector fields with 1 I = —— e,- A4.22)
fii 3 ... e 252 Variance of the spectral numbers (for some choice of the square roots). The matrix det(g(i>,-, Vj)) = 1 is constant as is the corresponding matrix for the flat vector fields. Therefore constant ¦ / = i — det(#op) 5 A4.23) Here H — 53 Vj о и,- is the socle field. One of the origins of the first part log r/ is the geometry of isomonodromic deformations. The second structure connections and the first structure connec- connections of the semisimple Frobenius manifold are isomonodromic deformations overP1 x M of restrictions to a slice P1 x {r}. The function Г/ is their г -function in the sense of [JMMS][JMU][JM][Mal4]. See [Sab4] for other general ref- references on this. The situation for Frobenius manifolds is discussed and put into a Hamiltonian framework in [Du3, Lecture 3], [Man2, II§2], and in [Hi]. The coefficients #,¦ of the 1-form dlog T/ = ? Hjdiij are certain Hamiltonians and motivate the definition of this 1-form. Hitchin [Hi] compares the realizations of this for the first and the second structure connections. Another origin of the whole G-function comes from quantum cohomology. Getzler [Ge] studied the relations between cycles in the moduli space M. \ ,4 and derived from it recursion relations for genus one Gromov-Witten invariants of projective manifolds and differential equations for the genus one Gromov- Witten potential. Dubrovin and Zhang [DuZl, chapter 6] investigated these differential equa- equations for any semisimple Frobenius manifold and found that they have always one unique solution (up to addition of a constant), the G-function, a function which was proposed in [Gi7]. They also proved the major part of the conjectures in [Gi7] concerning G(t). Finally, they found that the potential of the Frobenius manifold (for genus zero) and the G-function (for genus one) are the basements of full free energies in genus zero and one and give rise to Virasoro constraints [DuZ2]. Givental found formulas for the full free energies at higher genus [,Gi8] and proved Virasoro constraints for them [Gi9]. Exploiting this for singularities will be a big task for the future. For our application in section 13.3 we need only the definition of log T/ and the behaviour of G(t) with respect to the Euler field and the caustic in a massive Frobenius manifold. We have to summarize some known formulas related to the canonical coordinates of a semisimple Frobenius manifold ([Du3], [Man2], also [Gi7]). 14.2 G-function of a massive Frobenius manifold 253 The 1-form e = g{e,.) is closed and can be written as e = dr\. One defines j)i := ejT) - g{et,e) = g(eit e,), raj :=e,-e/4 = ejTjj =е}гц, Vjj :=-(«; -Uj)yij, l ¦ dlogr/ := - n A4.24) A4.25) A4.26) A4.27) A4.28) A4.29) Theorem 14.4 Let (M, o,e, E, g) be a semisimple Frobenius manifold with global canonical coordinates u\, ...,um. (a) The rotation coefficients Yij (for i Ф j) satisfy the Darboux-Egorojf equations ekyij = MYkj &гкф1ф]фк, A4.30) еуц = 0 fori^j. A4.31) (b) The connection matrix of the flat connection for the basis v\, ...,vm from A4.22) is the matrix {yi}d{uj - uj)). The Darboux-Egorojf equations are equivalent to the flatness condition d(yud(Ui - uj)) + (y,7d(«; - uj)) л (Yijd(uj - Uj)) = 0. A4.32) (c) The 1-form d log Г/ is closed and comes from a function log r,. (d) The Euler field E satisfies Е(гц) = (D - 2)гц and E{yu) = -yu. (e) If the canonical coordinates are chosen such that E = ? u'e' tlten tlte matrix -(V,j) is the matrix of the endomorphism § i T ih respect to the basis V\, ..., vm. VE-§ id on TM with Proof, (a) and (b) See [Du3, pp. 200-201] or [Man2, I§3]. (c) This can be checked easily with the Darboux-Egoroff equations. (d) It follows from Lie?(g) = D ¦ g and from [e,-, E] = e,. (e) This is implicit in [Du3, pp. 200-201]. One can check it with (a) and (b) and(d). D The endomorphism V is skewsymmetric with respect to g and flat with eigen- eigenvalues d\ - f; the numbers dt can be ordered such that d\ = 1, dj + dm+1 _,• = D (cf. Remark 9.2 (e)).
Т'П пт- пт"il 1.. 32.' 3" "п.. ¦ '' :,|i !; 1 Vj i ' ¦ 254 Variance of the spectral numbers Corollary 14.5 [DuZl, Theorem 3] Suppose that E — J2ще-,. Then ? log т/ =-i 48 Proof: A4.33) =: у. A4.34) A4.35) Formula A4.23) shows E(J) = , tion of the G-function. i^=2 J. Now A4.34) follows from the defini- D If M is a massive Frobenius manifold with caustic /C, one may ask which kind of poles the 1 -form d log t/ has along /C and when the G-function extends over 1С. In [DuZl, chapter 6] the G-function is calculated for the 2-dimensional Frobenius manifolds h{n), n > 3, on M — C2 with coordinates (t\, t2) and e = —-. It turns out to be 1 B-иХЗ-и) A4.36) Especially, for the case /гC) = A2 the G-function is G(t) = 0. This was also checked in [Gi7]. Givental concluded that in the case of singularities the G-functionof thebase space of a semiuniversal unfolding with some Frobenius manifold structure extends holomorphically over the caustic. This is a good guess, but it does not follow from the case Аг, because a Frobenius manifold structure on a germ of an F-manifold of type A2^T~2 for ш > 3 is never the product of the Frobenius manifolds Аг and A™~2 (the numbers d\,..., dm would not be symmetric). However, it is true, as the following result shows. Theorem 14.6 Let (M, о, е, Е, g) be a simply connected massive Frobenius manifold. Suppose that at generic points of the caustic K. the germ of the underlying F-manifold is of type l2(n)A™~2 for one fixed number n > 3. 14.2 G-function of a massive Frobenius manifold 255 (a) The form dlog Г/ has a logarithmic pole along /C with residue — (62[- along K.reg. (b) The G-function extends holomorphically over K, if and only ifn = 3. Proof. Theorem 14.3 and A4.23) say that the form — ^d log J has a logarithmic pole along/C with residue ^ along ICreg. This equals ^f^f- if and only ifn = 3. So (b) follows from (a). It is sufficient to show (a) for the F-manifold M = Cm in the proof of Theorem 14.3, equipped with some metric which makes a Frobenius manifold out of it (we do not need an Euler field here). Unfortunately we do not have an identity for dlog Г/ as A4.7) for the socle field which would allow only a most convenient metric to be considered. We use A4.11)-A4.17) and A4.24)-A4.29) and consider a neighbourhood of 0 6 Cm = M. Denote for j > 3 П% Ъ T2J := («i - «y)-^- - ( - иу) n2 A4.37) With ) ф 0 for j > 3, A4.29) and A4.16) one calculates 8dlogt/ — holomorphic 1-form + T\2 tnf + ? T2jtfdt2 - Y, Tijduj. A4.38) From A4.14) one obtains 1 у>з 2 у>з 1 -n+21 1 n-2 4 2 A4.39) A4.40) A4.41) The vector 52|o is a generator of the socle of the subalgebra in TqM which corresponds to I2(n). Therefore 52(^)@) # 0. It is not hard to see with A4.39)
1 'Г! Я 256 Variance of the spectral numbers 257 and A4.16) that the terms T{j and T2j t2 2 for j > 3 are holomorphic at 0. The term 7"i2 is T\2 = - • = 8 . n • 1- holomorphic 1-form. A4.42) In t2 This proves part (a). ? Remarks 14.7 (a) The base spaces of semiuniversal unfoldings meet the case n = 3 in Theorem 14.6. Closely related are base spaces of certain unfoldings of tame functions. The germs of F-manifolds are isomorphic to products of the germs of F-manifolds from hypersurface singularities. But Sabbah [Sab3] [Sab2][Sab4] equipped them with a metric such that the Frobenius manifold structure is in general not a product (cf. section 11.4). (b) It might be interesting to look for massive Frobenius manifolds which meet the case n = 3 in Theorem 14.6, but where the underlying F-manifolds are not locally products of those from hypersurface singularities. In view of Theorem 5.6 the analytic spectrum of such F-manifolds would have singulari- singularities, but only in codimension > 2, as the analytic spectrum of Аг is smooth. The analytic spectrum is Cohen-Macaulay and even Gorenstein and a Lagrange variety. P. Seidel (Ecole Polytechnique) showed me a normal and Cohen-Macaulay Lagrange surface. But it seems to be unclear whether there exist normal and Gorenstein Lagrange varieties which are not smooth. 14.3 Variance of the spectrum By Theorem 14.6 the germ (M, 0) of a Frobenius manifold as in Theorem 11.1 for an isolated hypersurface singularity / has a holomorphic G-function G(t), unique up to addition of a constant. By Corollary 14.5 and Theorem 11.1 this function satisfies III ГУ .. ГУ1 1 y. A4.43) -a,) 48 So it has a very peculiar strength: it gives a hold at the squares of the spectral numbers ai,..., aM of the singularity. Because of the symmetry a,- +aM+i_/ = n — 1, the spectral numbers are scattered around their expectation value '¦—-. One may ask about their variance - Yli=\(.ai ~ ^y^J- 14.3 Variance of the spectrum Conjecture 14.8 The variance of the spectral numbers of an isolated hyper- hypersurface singularity is or, equivalently, -I Y > Theorem 14.9 In the case of a quasihomogeneous singularity f n-Г2 and y=0. A4.44) A4.45) A4.46) A4.47) Proof. One has the isomorphism (G/J/, mult., [/]) = (TqM, o, E\q). Here / € Jf and ?|o = 0 and therefore E G(t) =0. ? Remarks 14.10 (a) When I presented Theorem 14.9 at the summer school on singularity theory in Cambridge in August 2000, I asked for an elementary proof of it. This was found by A. Dimca. It uses the characteristic function of the spectral numbers. The variance is A4.48) A4.49) In the case of a quasihomogeneous singularity with weights wo, ...,wn 6 @, |] and degree 1 the characteristic function is A4.50) as is well known. Using this product formula A. Dimca [Di] showed that in the case of a quasihomogeneous singularity the variance is ?"_0 ^-yj^ = a"l~2"' ¦ He also made a conjecture dual to Conjecture 14.8 forthe case of tame polynomials: there the inverse inequality to A4.44) should hold. The conjectures intersect in the case of quasihomogeneous singularities and give there the equality A4.46).
f m 'IP!' ¦ 258 Vfan'ance of the spectral numbers (b) M. Saito in September 2000 proved Conjecture 14.8 in the case of irre- irreducible plane curve singularities [SM5]. (c) Т. Brelivet in May 2001 proved it in the case of plane curve singularities with nondegenerate Newton polyeder [Bre]. (d) The only unimodal or bimodal families of not semiquasihomogeneous singularities are the cusp singularities Tpqr and the 8 bimodal series. The spectral numbers are given in [AGV2]. One finds with equality only for the simple elliptic singularities. In the case of the 8 bimodal families one obtains У = -rf~ • (l j—) > 0 A4.52) 48-к \ p + kJ with к :— 9, 7, 6, 6, 5 for ?з,р, Z\ p, Q2,P, W\p, S\iP, respectively, and Y = 48-at A4-53> with к := 6, 5, § for w\ p, S\p, U\tP, respectively. (e) At the summer school in Cambridge in August 2000 Conjecture 14.8 was confirmed for many other singularities using the computer algebra system Singular and especially the program of M. Schulze for computing spectral numbers, which is presented in [SchuSt]. (f) In [SK8] K. Saito studied the distribution of the spectral numbers and their characteristic function Xf heuristically and formulated several questions about them. The G-function might help these problems to be continued. (g) One can speculate that Conjecture 14.8, if it is true, comes from a more profound hidden interrelation between the GauB-Manin connection and po- polarized mixed Hodge structures. The existence of Frobenius manifolds and G-functions alone is not sufficient, as the following shows. In Remark 11.7 (b) an example of M. Saito [SM3,4.4] is sketched which leads for the semiquasihomogeneous singularity / = x6 + y6 + x4y4 to Frobenius manifold structures with {d\,.., d^} ф {1+ai—a,- \i = 1,..., /x}. The number у in that case is у = — ¦— < 0. (h) In the case of the simple singularities Ak, Дь E$, E-j, E$, the parame- parameters 11,..., ?M of a suitably chosen unfolding are weighted homogeneous with positive degrees with respect to the Euler field. Therefore G = 0 in these cases (cf. [Gi7]) 14.3 Variance of the spectrum Lemma 14.11 The number у of the sum f(xo,..., x,,) + g(yo, ¦ singularities f and g satisfies y(f + g) = y(g) ¦ y(f). 259 ¦, Ут) of two A4.54) Proof: Let ai,..., or^/) and /3i,..., /3M<g) denote the spectral numbers of / and g. Then the spectrum of / + g as an unordered tuple is [AGV2][SchSt] (a,- + pj + 1 | i = 1,..., fj.(f), 7 = 1,..., ii(g)). A4.55) This and the symmetry of the spectra yields A4.54). D Remarks 14.12 For any Frobenius manifold the variance ~ Ya=Mi ~ fJ of the eigenvalues d\,...,dm of V? is interesting. It turns up not only as in Corollary 14.5 related to the G-function in the semisimple case, but also in the operator Lq of the Virasoro constraints in [DuZ2, B.30)] for any Frobenius manifold. Prior to [DuZ2] the Virasoro constraints were postulated in the case of quan- quantum cohomology of projective manifolds with hp-q = 0 for p ф q in [EHX]. There a formula for the variance was considered which turned out to be a special case of the following formula from [LiW] (cf. also [Bori]), which is valid for any projective manifold: p.q A4.56) Here c\ is the /th Chern class of the manifold, n is its dimension. The proof uses the Hirzebruch—Riemann—Roch theorem. The formula is generalized to projective varieties with at most Gorenstein canonical singularities in [Bat]. Comparing the right hand side with the singularity case, one can speculate n ~ aM — ori, с„ ~ ц. and ask about ^с\сп_\ ~?.
Bibliography 261 Bibliography [AB] Anosov, D. V., A. A. Bolibmch: The Riemann-Hilbert problem. Aspects of mathematics, vol. 22, Vieweg, Braunschweig Wiesbaden 1994. [Arl] Arnold, V. I.: Critical points of functions and the classification of caustics (in Russian). Uspekhi Mat. Nauk 29.3 A974), 243-244. [Ar2] Arnold, V. I.: Critical points of functions on a manifold with boundary, the simple Lie groups Bk, Q, and F4 and singularities of evolutes. Russian Math. Surveys 33.5 A978), 99-116. [АгЗ] Arnold, V. I.: Indices of singular points of 1-forms on a manifold with boundary, convolution of invariants of reflection groups, and singular projections of smooth surfaces, Russian Math. Surveys 34.2 A979), 1—42. [AGV1] Arnold, V. I., S. M. Gusein-Zade, A. N. Varchenko: Singularities of differentiable maps, volume I. Birkhauser, Boston 1985. [AGV2] Arnold, V. I., S. M. Gusein-Zade, A. N. Varchenko: Singularities of differentiable maps, volume II. Birkhauser, Boston 1988. [Au] Audin, M.: Symplectic geometry in Frobenius manifolds and quantum cohomology. J. of Geometry and Physics 25 A998), 183-204. [Bal] Barannikov, S.: Generalized periods and mirror symmetry in dimensions n > 3. Preprint, math.AG/9903124. [Ba2] Barannikov, S.: Quantum periods -1. Semi-infinite variations of Hodge structures. Preprint, math.AG/0006193. [Ba3] Barannikov, S.: Semi-infinite Hodge structures and mirror symmetry for projective spaces. Preprint, math.AG/0010157. [BaK] Barannikov, S., M. Kontsevich: Frobenius manifolds and formality of Lie algebras of polyvector fields. Int. Math. Res. Notices 1998-4, 201-215. [Bat] Batyrev, V. V.: Stringy Hodge numbers and Virasoro algebra. Preprint, alg-geom/9711019. [Bj] Bjork, J.-E.: Analytic ©-modules and applications. Kluwer Academic Publishers, Mathematics and its applications vol. ^47, Dordrecht Boston London 1993. [Bore] Borel, A.: Introduction aux groupes arithmetiques. Hermann, Paris 1969. [Bori] Borisov, L. A.: On the Betti numbers and Chern classes of varieties with trivial odd cohomology groups. Preprint, alg-geom/9703023. [Bou] Bourbaki, N.: Groupes et algebres de Lie, chapitres 4, 5 et 6. Hermann, Paris 1968. [Bre] Brelivet, Т.: Variance of the spectral numbers and Newton polygons. Preprint, May 2001. [Bril ] Brieskorn, E.: Singular elements of semisimple algebraic groups. In: Actes Congres Int. Math. 2, Nice A970), 279-284. [Bri2] Brieskorn, E.: Die Monodromie der isolierten Singularitaten von Hyperflachen. Manuscripta Math. 2 A970), 103-161. [BrL] Brocker, Th., L. Lander: Differentiable germs and catastrophes. London Math. Soc. Lecture Note Series 17, Cambridge University Press 1975. [Bru] Bruce, J. W.: Vector fields on discriminants and bifurcation varieties. Bull. London Math. Soc. 17 A985), 257-262. [CDGP] Candelas, P., X. С de la Ossa, P. S. Green, L. Parkes: A pair of Calabi-Yau maifolds as an exactly soluble superconformal field theory. Nucl. Phys. B359 A991), 21-74. [Ca] Cartan, H.: Quotient d'un espace analytique par un groupe d'automorphismes. In: Algebraic geometry and topology, a symposium in honour of S. Lefschetz. Princeton Univ. Press, 1957, 90-102. [CaK] Cattani, E., A. Kaplan: Polarized mixed Hodge structures and the local monodromy of a variation of Hodge structure. Invent. Math. 67 A982), 101-115. [ChV] Chmutov, S. V, A. N. Varchenko: Tangent cone to the д = constant stratum in terms of mixed Hodge structure. Moscow Univ. Math. Bull. 40 A985), 7-12. [Co] Coxeter, H. S. M.: Discrete groups generated by reflections. Ann. Math. 35 A934), 588-621. [DD] Nguyen tien Dai, Nguyen huu Due: Stabilite de l'interaction geometrique entre deux composantes holonomes simples. С R. Acad. Sc. Paris, Serie A, 291A980), 113-116. [Del] Deligne, P.: Equations differentielles a points singuliers reguliers. Lecture Notes in Math. vol. 163, Springer, Berlin Heidelberg New York 1970. [De2] Deligne, P.: Theorie de Hodge, II. Publ. Math. I.H.E.S. 40 A971), 5-57. [DVV] Dijkgraaf, R., E. Verlinde, H. Verlinde: Topological strings in d < 1. Nucl. Phys. В 352 A991), 59-86. [Di] Dimca, A.: Monodromy and Hodge theory of regular functions. Proceedings of the summer school on singularities, Newton Institute, Cambridge 2000, to appear. [Dul] Dubrovin, В.: Integrable systems in topological field theory. Nucl. Phys. В 379A992),627-689. [Du2] Dubrovin, В.: Differential geometry of the space of orbits of a Coxeter group. In: Surveys in differential geometry: integral systems [integrable systems]. Surveys in Diff. Geom. IV (C. L. Terng, K. Uhlenbeck, eds.). Int. Press, Boston MA 1998, 181-211. [Du3] Dubrovin, В.: Geometry of 2D topological field theories. In: Integrable sytems and quantum groups. Montecatini, Terme 1993 (M Francoviglia, S. Greco, eds.). Lecture Notes in Math. 1620, Springer-Verlag 1996, 120-348. [Du4] Dubrovin, В.: Painleve transcendents in two-dimensional topological field theory. In: The Painleve property. One century later (ed. Roberto Conte.). CRM Sen Math. Phys.. Springer, New York 1999, 287-412. [Du5] Dubrovin, В.: Geometry and analytic theory of Frobenius manifolds. In: Proceedings of the ICM 1998, Vol II, pp. 315-326. [DuZl] Dubrovin, В., Y. Zhang: Bihamiltonian hierarchies in 2D topological field theory at one-loop approximation. Commun. Math. Phys. 198 A998), 311-361. 260
262 Bibliography [DuZ2] Dubrovin, В., Y. Zhang: Frobenius manifolds and Virasoro constraints. Sel. math., New ser. 5 A999), 423-466. [EHX] Eguchi, Т., К. Hori, Ch.-Sh. Xiong: Quantum cohomology and Virasoro algebra. Phys. Lett. B402 A997), 71-80. [EV] Esnault, H., Viehweg, E.: Logarithmic de Rham complexes and vanishing theorems. Invent. Math. 86 A986), 161-194. [Fi] Fischer, G.: Complex analytic geometry. Lecture Notes in Math. vol. 538, Springer, Berlin Heidelberg New York 1976. [Ga] Gabrielov, A. M.: Bifurcations, Dynkin diagrams and modality of isolated singularities. Funct. Anal. 8 A974),_94-98. [Ge] Getzler, E.: Intersection theory on M.\^ and elliptic Gromov-Witten invariants. J. Amer. Math. Soc. 10 A997), 973-998. [Gil] Givental, А. В.: Convolution of invariants of groups generated by reflections, and connections with simple singularities of functions. Funct Anal. 14A980), 81-89. [Gi2] Givental, А. В.: Singular Lagrangian manifolds and their Lagrangian maps. J. Soviet Math. 52.4 A988), 3246-3278. [Gi3] Givental, А. В.: Homological geometry and mirror symmetry. In: Proceedings of the ICM-94 Zurich. Birkhauser, Basel 1995, 472-480. [Gi4] Givental, А. В.: Equivariant Gromov-Witten invariants. Int. Math. Res. Notices 1996-13, 613-663. [Gi5] Givental, А. В.: A mirror theorem for tone complete intersections. In: Topological field theory, primitive forms and related topics (Kyoto, 1996). Progr. in Math. 160, Birkhauser, Boston MA 1998, 141-175. [Gi6] Givental, А. В.: Stationary phase integrals, quantum Toda lattices, flag manifolds and the mirror conjecture. In: Topics in singularity theory. Amer. Math. Soc. Transl. Ser. 2,180, Amer. Math. Soc., Providence RI 1997, 103-115. [Gi7] Givental, А. В.: Elliptic Gromov-Witten invariants and the generalized mirror conjecture. In: Integrable systems and algebraic geometry. Proceedings of the Taniguchi Symposium 1997 (M.-H. Saito, Y. Shimizu, K. Ueno, eds.). World Scientific, River Edge NJ 1998, 107-155. [Gi8] Givental, А. В.: Semisimple Frobenius structures at higher genus. Preprint, math.AG/0008067. [Gi9] Givental, А. В.: Gromov-Witten invariants and quantization of quadratic hamiltonians. Preprint, math.AG/0108100. [Gre] Greuel, G.-M.: Der GauB-Manin-Zusammenhang isolierter Singularitaten von vollstandigen Durchschnitten. Math. Ann. 214 A975), 235-266. [GrHP] Greuel, G.-M., С Hertling, G. Pfister: Moduli spaces of semiquasihomogeneous singularities with fixed principal part. J. Algebraic Geometry 6 A997), 169-199. [Gri] Griffiths, P.: Periods of integrals on algebraic manifolds: summary of main results and open problems. Bull. Amer. Math. Soc. 76 A970), 228-296. [GrH] Griffiths, P., J. Harris: Principles of algebraic geometry. John Wiley and sons, New York 1978. [Hel] Hertling, C: Analytische Invarianten bei den unimodularen und bimodularen Hyperflachensingularitaten. Dissertation. Bonner Math. Schriften 250 , Bonn 1992. [He2] Hertling, C: Ein Torellisatz fur die unimodalen und bimodularen Hyperflachensingularitaten. Math. Ann. 302 A995), 359-394. Bibliography 263 [He3] Hertling, C: Brieskorn lattices and Torelli type theorems for cubics in P3 and for Brieskorn-Pham singularities with coprime exponents. In: Singularities, the Brieskorn anniversary volume. Progress in Mathematics 162. Birkhauser Verlag, Basel Boston Berlin 1998, 167-194. [He4] Hertling, C: Classifying spaces and moduli spaces for polarized mixed Hodge structures and for Brieskorn lattices. Compositio Math. 116 A999), 1-37. [HM] Hertling, C, Yu. Manin: Weak Frobenius manifolds. Int. Math. Res. Notices 1999-6, 277-286. [Hi] Hitchin, N. J.: Frobenius manifolds (notes by D. Calderbank). In: Gauge Theory and symplectic geometry, Montreal 1995, J. Hurtubise and F. Lalonde eds., Kluwer Academic Publishers, Netherlands 1997, 69-112. [Hoe] Hormander, L.: Fourier integral operators I. Acta. Math. 127 A971), 71-183. [Hoi] Holmann, H.: Komplexe Raume mit komplexen Transformationsgruppen. Math. Ann. 150 A963), 327-360. [Hou] Houzel, C: Geometrie analytique locale, II, theorie des morphismes finis. Seminaire Henri Cartan 13e annee, no. 19, 1960/61. [JM] Jimbo, M., T. Miwa: Monodromy preserving deformations of linear ordinary differential equations with rational coefficients II. Physica 2D A981),407-448. [JMMS] Jimbo, M., T. Miwa, Y. Mori, M. Sato: Density matrix of impenetrable Bose gas and the fifth Painleve transcendent. Physica ID A980), 80-158. [JMU] Jimbo, M., T. Miwa, K. Ueno: Monodromy preserving deformations of linear ordinary differential equations with rational coefficients I. Physica 2D A981), 306-352. [Kar] Karpishpan, Y: Torelli theorems for singularities. Invent. Math. 100 A990), 97-141. Erratum. Invent. Math. 108 A992), 667-668. [Kasl] Kashiwara, M.: Quasi-unipotent constructible sheaves. J. Fac. Sci. Univ. Tokyo, Sec. IA 28 A982), 757-773. [Kas2] Kashiwara, M.: Introduction to microlocal analysis. Enseign. Math. 32 A986), 227-259. [KK] Kashiwara, M., T. Kawai: On holonomic systems of microdifferential equations III - systems with regular singularities. Publ. RIMS, Kyoto Univ. 17A981)813-979. [Каир] Каир, L., В. Каир: Holomorphic functions of several variables. De Gruyter, studies in mathematics 3, Berlin New York 1983. [KM] Kontsevich, M., Yu. Manin: Gromov-Witten classes, quantum cohomology, and enumerative geometry. Commun. Math. Phys. 164 A994), 525-562. [Ku] Kulikov, Va.S.: Mixed Hodge structures and singularities. Cambridge tracts in mathematics 132, Cambridge University Press, 1998. [Кип] Kunz, E.: Komplementarmodul und kanonischer Modul. In: Der kanonische Modul eines Cohen-Macaulay-Rings (Herzog, J., Kunz, E., eds.). Lecture Notes in Math. vol. 238, Springer, Berlin Heidelberg New York 1973, 85-102. [La] Lazzeri, F.: A theorem on the monodromy of isolated singularities. Singularites a Cargese, Asterisque 7-8 A973), 269-275. [Le] Le Dung Trang: Une application d'un theoreme d'A'Campo a l'equisingularite. Preprint, Centre de Math, de l'Ecole Polytechnique, Paris, 1973.
264 Bibliography [LeR] Le Dung Trang, C. P. Ramanujam: The invariance of Milnor's number implies the invariance of the topological type. Amer. J. of Math. 98 A973), 67-78. [LiW] Libgober, A. S.. J. W. Wood: Uniqueness of the complex structure on Kahler manifolds of certain homotopy types. J. Diff. Geom. 32 A990), 139-154. [Lol ] Looijenga, E. J. N: The complement of the bifurcation variety of a simple singularity. Invent Math. 23 A974), 105-116. [Lo2] Looijenga, E. J. N.: Isolated singular points on complete intersections. London Math. Soc. Lecture Note Series 77, Cambridge University Press 1984. [Lu] Luengo, I.: The д-constant stratum is not smooth. Invent. Math. 90 A987), 139-152. [Lyl] Lyashko, O. V: The geometry of bifurcation diagrams. Russian Math. Surveys 34.3 A979), 209-210. [Ly2] Lyashko, O. V: Classification of critical points of functions on a manifold with singular boundary. Funct. Anal. 17.3A983), 187-193. [Ly3] Lyashko, O. V: Geometry of bifurcation diagrams. J. Soviet Math. 27 A984), 2736-2759. [Mall] Malgrange, В.: Integrates asymptotiques et monodromie. Ann. Sci. Ecole Norm. Sup. 7 A974), 405-430. [Mal2] Malgrange, В.: Modules microdifferentiels et classes de Gevrey. In: Mathematical analysis and applications, essays dedicated to L. Schwartz on the occasion of his 65th birthday. Adv. in Math. Suppl. Studies 7B, Academic Press Inc. 1981, 515-530. [Mal3] Malgrange, В.: Deformations de systemes differentiels et microdifferentiels. In: Seminaire de'l ENS, Mathematique et Physique, 1979-1982, Progress in Mathematics vol. 37, Birkhauser, Boston 1983, pp. 353-379. [Mal4] Malgrange, В.: Sur les deformations isomonodromiques, I, II. In: Seminaire de'l ENS, Mathematique et Physique, 1979-1982, Progress in Mathematics vol. 37, Birkhauser, Boston 1983, pp. 401-438. [Mal5] Malgrange, В.: Deformations of differential systems, II. J. Ramanujan Math. Soc. 1A986), 3-15. [Mal6] Malgrange, В.: Connexions meromorphes. In: Congres Singularites, Lille 1991 (P. Brasselet ed.), Cambridge University Press, 1994, pp. 251-261. [Mal7] Malgrange, В.: Connexions meromorphes II: le reseau canonique. Invent. Math. 124 A996), 367-387. [Manl] Manin, Yu.: Moduli Fuchsiani. Annali Scuola Normale Sup. di Pisa Sen III 19A965), 113-126. [Man2] Manin, Yu.: Frobenius manifolds, quantum cohomology, and moduli spaces. American Math. Society, Colloquium Publ. v. 47, 1999. [MaM] Manin, Yu., S. A. Merkulov: Semisitnple Frobenius (super)manifolds and quantum cohomology of P'. Topological methods in nonlinear analysis, Journal of the Juliusz Schauder Center 9 A997), 107-161. [Mathl] Mather, J.: Stability of C°°-maps III. Finitely determined map germs. Publ. Math. I.H.E.S. 35 A968), 127-156. [Math2] Mather, J.: Stability of C°°-maps IV. Classification of stable germs by K-algebras. Publ. Math. I.H.E.S. 37 A969), 223-248. [Matt] Mattel, J.-F.: Modules de feuilletages holomorphes singuliers: I equisingularite. Invent. Math. 103 A991), 297-325. [Mi] Milnor, J.: Singular points of complex hypersurfaces. Ann. Math. Stud. vol. 61, Princeton University Press 1968. Bibliography 265 [Muel] Miiller, G.: Reduktive Automorphismengruppen analytischer C-Algebren. J. Reine angew. Math. 364 A986), 26-34. [Mue2] Miiller, G.: Symmetry groups of (semi-)weighted homogeneous singularities. Preprint 1996. [Nam] Namikawa, Y.: Higher residues associated with an isolated hypersurface singularity. In: Algebraic varieties and analytic varieties, Advanced Studies in Pure Math. 1, North-Holland Publ. Company, 1983, pp. 181-193. [Nav] Navarro-Aznar, V.: Sobre la invariancia topologica de la multiplicitat. Publ. Sec. Mat. Univ. Autonoma Barcelona 20 A980), 261-262. [NS] Nemethi, A., C. Sabbah: Semicontinuity of the spectrum at infinity. Abh. math. Sem. Univ. Hamburg 69 A999), 25-35. [NZ] Nemethi, A., A. Zaharia: On the bifurcation set of a polynomial function and Newton boundary. Publ. RIMS, Kyoto Univ. 26 A990), 681-689. [No] Noumi, M.: Regular holonomic systems and their minimal extensions I. In: Group representations and systems of differential equations, Advanced Studies in Pure Math. 4, North-Holland Publ. Company, 1984, 209-221. [Odl] Oda, T: Introduction to algebraic analysis on complex manifolds. In: Algebraic varieties and analytic varieties, Advanced Studies in Pure Math. 1, North-Holland Publ. Company, 1983, 29-54. [Od2] Oda, Т.: К. Saito's period map for holomorphic functions with isolated singularities. In: Algebraic geometry, Sendai 1985, Advanced Studies in Pure Math. 10, 1987, 591-648. [Phi] Pham, F: Singularites des systemes differentiels de Gauss-Manin. Progress in Math. 2, Birkhauser, Boston 1979. [Ph2] Pham, F.: Remarque sur l'equivalence des fonctions de phase. С R. Acad. Sc. Paris, Serie A 290 A980), 1095-1097. [Ph3] Pham, F.: Structures de Hodge mixtes associees a un germe de fonction a point critique isole. Asterisque 101-102 A983), 268-285. [Ph4] Pham, R: Vanishing homologies and the n variable saddlepoint method. In: Singularities, Proc. of symp. in pure math. 40.2 A983), 319-333. [Ph5] Pham, R: La descente des cols par les onglets de Lefschetz, avec vues sur Gauss-Manin. Systemes differentiels et singularites. Asterisques 130 A985), 11-47. [Sabl] Sabbah, C: Monodromy at infinity and Fourier transform. Publ. RIMS, Kyoto Univ. 33 A997), 643-685. [Sab2] Sabbah, C: Frobenius manifolds: isomonodromic deformations and infinitesimal period mappings. Expo. Math. 16 A998), 1-58. [Sab3] Sabbah, C: Hypergeometric period for a tame polynomial. C. R. Acad. Sci. Paris Ser. I Math. 328 A999), 603-608, and D2 pages) math.AG/9805077. [Sab4] Sabbah, C.: Deformations isomonodromiques et varietes de Frobenius, une introduction. Centre de Mathematiques, Ecole Polytechnique, U.M.R. 7640 du C.N.R.S., no. 2000-05, 251 pages. [SKI] Saito, K.: Quasihomogene isolierte Singularitaten von Hyperflachen. Invent. Math. 14 A971), 123-142. [SK2] Saito, K.: Regularity of Gauss-Manin connection of flat family of isolated singularities. In: Quelques journees singulieres, Ecole Polytechnique, Paris, 1973. [SK3] Saito, K.: On a linear structure of the quotient variety by a finite reflexion group. Preprint RIMS-288 A979), Publ. RIMS, Kyoto Univ. 29 A993), 535-579.
1 266 Bibliography [SK4] Saito, К.: Theory of logarithmic differential forms and logarithmic vector fields, J. Fac. Sci. Univ. Tokyo, Sec. IA 27 A981), 265-291. [SK5] Saito, K.: A characterisation of the intersection form of a Milnor's fiber for a function with an isolated critical point. [SK6] Saito, K.: Primitive forms for a universal unfolding of a function with an isolated critical point, J. Fac. Sci. Univ. Tokyo, Sect. IA Math. 28 A982), 775-792. [SK7] Saito, K.: The higher residue pairings K^ for a family of hypersurface singular points. In: Singularities, Proc. of symp. in pure math. 40.2 A983), 441-463. [SK8] Saito, K.: The zeroes of characteristic function x/ f°r the exponents of a hypersurface isolated singular point. In: Algebraic varieties and analytic varieties, Advanced Studies in Pure Math. 1, North-Holland Publ. Company, 1983, 195-217. [SK9] Saito, K.: Period mapping associated to a primitive form. Publ. RIMS, Kyoto Univ. 19 A983), 1231-1264. [SMI] Saito, M.: Gauss-Manin systems and mixed Hodge structure, Proc. Japan Acad. 58 A982), 29-32. Supplement in Asterisque 101-102 A983), 320-321. [SM2] Saito, M.: On the structure of Brieskorn lattices. Preprint at Grenoble, September 1983. [SM3] Saito, M.: On the structure of Brieskorn lattices. Ann. Inst. Fourier Grenoble 39 A989), 27-72. [SM4] Saito, M.: Period mapping via Brieskorn modules. Bull. Soc. math. France 119A991), 141-171. [SM5] Saito, M.: Exponents of an irreducible plane curve singularity. Preprint, math. AG/0009133. [Schel] Scherk, J.: On the monodromy theorem for isolated hypersurface singularities. Invent. Math. 58 A980), 289-301. [Sche2] Scherk, J.: A propos d'un theoreme de Mather et Yau. C. R. Acad. Sci. Paris, Serie I, 296 A983), 513-515. [SchSt] Scherk, J., J. H. M. Steenbrink: On the mixed Hodge structure on the cohomology of the Milnor fibre. Math. Ann. 271 A985), 641-665. [Schm] Schmid, W.: Variation of Hodge structure: The singularities of the period mapping. Invent. Math. 22 A973), 211-319. [SchuSt] Schulze, M., J. H. M Steenbrink: Computing Hodge-theoretic invariants of singularities. Proceedings of the summer school on singularities, Newton Institute, Cambridge 2000, to appear. [Seb] Sebastiani, M.: Preuve d'une conjecture de Brieskorn. Manuscripta Math. 2 A970), 301-308. [Ser] Serre, J.-R: Prolongement de faisceaux analytiques coherents. Ann. Inst. Fourier, Grenoble 16 A966), 363-374. [ShSI] Shcherbak, I., A. Szpirglas: Boundary singularities: Topology and duality. In: Advances in Soviet Math. 21, 1994, 213-223. [ShS2] Shcherbak, I., A. Szpirglas: Boundary singularities: double coverings and Picard-Lefschetz formulas. С R. Acad. S. Paris, Serie I, 322 A996), 557-562. [ShO] Shcherbak, O. P.: Wavefronts and reflection groups. Russian Math. Surveys 43S3 A988), 149-194. [SI] Slodowy, P.: Einige Bemerkungen zur Entfaltung symmetrischer Funktionen. Math. Z. 158 A978), 157-170. Bibliography 267 [Stn] Steenbrink, J. H. M.: Mixed Hodge structure on the vanishing cohomology. In: Real and complex singularities, Oslo 1976, P. Holm (ed.). Alphen aan den Rijn: Sijthoff and Noordhoff 1977, 525-562. [Stv] Stevens, J.: On the д-constant stratum and the V-filtration: an example. Math. Z. 201A989), 139-144. [Sz] Szpirglas, A.: Singularites de bord: dualite, formules de Picard Lefschetz relatives et diagrammes de Dynkin. Bull. Soc. math. France 118 A990), 451-486. [Tel] Teissier, В.: Deformations a type topologique constant. Seminaire de geometrie analytique 1971-1972, Asterisque 16, 215-249. [Te2] Teissier, В.: The hunting of invariants in the geometry of discriminants. In: Real and complex singularities, Oslo 1976, P. Holm ed., Alphen a/d Rijn: Sijthoff & Noordhoff, 565-677. [Ter] Terao, H.: The bifurcation set and logarithmic vector fields. Math. Ann. 263 A983), 313-321. [Val] Varchenko, A. N.: The asymptotics of holomorphic forms determine a mixed Hodge structure. Sov. Math. Dokl. 22 A980), 112-115. [Va2] Varchenko, A. N.: Asymptotic mixed Hodge structure in vanishing cohomologies. Math. USSR Izvestija 18 A982), 469-512. [Va3] Varchenko, A. N.: The complex singular index does not change along the stratum д = constant. Functional Anal. Appl. 16 A982), 1-9. [Va4] Varchenko, A. N.: A lower bound for the codimension of the stratum ц = constant in terms of the mixed Hodge structure. Moscow Univ. Math. Bull. 37A982), 30-33. [Va5] Varchenko, A. N.: Local classification of volume forms in the presence of a hypersurface. Funct. Anal. 19.4 A985), 269-276. [Va6] Varchenko, A. N.: On the local residue and the intersection form on the vanishing cohomology. Math. USSR Izvestija 26 A986), 31-52. [VaS] Vassiliev, V. A., V. V. Serganova: On the number of real and complex moduli of singularities of smooth functions and realizations of matroids. Math. Notes 49 A991), 15-20. [Vo] Voisin, C: Symetrie miroir. Panoramas et syntheses 2, Soc. Math, de France 1996. [Wah] Wahl, J.: Equisingular deformations of plane algebroid curves. Trans. Amer. Math. Soc. 193 A974), 143-170. [Wall] Wall, С. Т. С: A note on symmetry of singularities. Bull. London Math. Soc. 12 A980), 169-175. [Wal2] Wall, С. Т. С: A second note on symmetry of singularities. Bull. London Math. Soc. 12 A980), 347-354. [War] Warner, F. W.: Foundations of differentiable manifolds and Lie groups. Graduate texts in mathematics 94. Springer-Verlag 1983. [Was] Wassermann, G.: Stability of unfoldings. Lecture Notes in Math. 393, Springer-Verlag, 1974. [Wir] Wirthmiiller, K.: Singularities determined by their discriminant. Math. Ann. 252 A980), 237-245. [Wit] Witten, E.: Two-dimensional gravity and intersection theory on moduli spaces. Surveys in Diff. Geom. 1 A991), 243-310. [Za] Zakalyukin, V. M.: Reconstruction of wave fronts depending on one parameter. Funct. Anal. 10.2 A976), 139-140. 1 к к
Index Н°°, ПО, 136, 189, 196 Н", 168 «№) ciVtt, 168 Пт, 172, 174 Щ, 171, 188, 191 д-constant family, 193 д-constant stratum, 53, 218 д-homotopy class, 241 т-function, 252 analytic spectrum, 11, 24,56 automorphism group, 52 bifurcation diagram, 36,161 Birkhoff problem, 120, 214,216 boundary singularity, 69 Brieskorn lattice, 114, 172, 188 canonical coordinates, 18, 160 canonical extension, 142 caustic, 13,36, 161 classifying space for Brieskorn lattices, 194 classifying space for PMHSs, 188 cohomology bundle, 168 covariant derivative, 20, 146 Coxeter group, 75,83 critical space, 62 Darboux-Egoroff equations, 253 de Rham cohomology, 174 deformed flat coordinates, 205 Deligne's /''¦«, 184 development, 30, 41, 246 discriminant, 36,40,47, 161 discriminant of a singularity, 66, 167 eigenspace, 111 eigenspace decomposition, 10 elementary part, 112 elementary section, 110, 136 Eulerfield, 14,25,29, 146 exhaustive filtration, 117 F-manifold, 14 filtration, 114, 115, 117 first structure connection, 154, 205 flat coordinates, 147 flat metric, 83 flat vector bundle, 109 Fourier-Laplace transformation, 156, 214,216 free divisor, 47, 134 Frobenius algebra, 10 Frobenius manifold, 22, 83, 146 front, 30 G-function, 251 GauB-Manin connection, 170 GauB-Manin system, 179, 181 Gelfand-Leray form, 171 generalized Milnor fibration, 168 generating family, 59, 68 generating function, 30, 36 good section, 122 Gorenstein ring, 10 Gromov-Witten invariants, 252 Grothendieck residue, 180 group of symmetries, 235 higher residue pairings, 180, 191 hypersurface singularity, 62, 165 infinitesimal Torelli type result, 226 intersection form, 150, 153, 181, 189 isolated hypersurface singularity, 62, 165 isomonodromic deformations, 252 269
, 62 i|iMiiiiiiii(iilu<in<>lii|iy, -II. 252 KcHlniru Spencer niiip, 62. 166 Uigmnyc tihraiion. 31 I.agrange map, 31 Lagrange variety, 24 lattice, I 13, 115 Lefschetz thimble, 215 Levi-Civita connection, 146 Lie derivative, 14, 146 logarithmic differential form, 131 logarithmic pole, 118, 134, 158, 162 logarithmic vector field, 47, 131 Lyashko-Looijenga map, 30, 36, 55, 80 M-tame function, 213, 217 massive, 24 massive Frobenius manifold, 160 metric, 145 microdifferential operator, 113, 174 Milnor fibration, 168 miniversal Lagrange map, 33 mirror symmetry, 211 mixed Hodge structure, 184, 193 modality, 53 moderate growth, 112 moduli of germs of F-manifolds, 93 moduli space Мц, 225, 241 monodromy, 110, 162, 189 monodromy group, 153, 168, 204 multiplication invariant, 10, 21, 146 multivalued section, 110 normal crossing case, 132, 140 open swallowtail, 95 opposite filtration, 122, 185, 197 order, 112 oscillating integral, 157, 214, 216 period map, 225 Picard-Lefschetz transformation, 153, 168 PMHS, 184 Poincare rank, 134 polarized mixed Hodge structure, 184, 192 polarizing form, 184 pole of order < r + 1, 134 potential, 22, 147 potentiality, 22, 146 primitive form, 104, 178, 202 primitive subspace, 183, 185 principal part, 112 reduced Kixlinni Spencer шар, 62, 163 reduced l.ynshko-l.<xiijcn};ii map, 30, M> rellexive, 133 reflexive extension, 133, 139 regular singular. 143 residual connection. 135. 137. 158 residue endoniorphism, 114, 135, 137, 158, 162 restricted bifurcation diagram, 38, 50, 51 restricted caustic, 38 restricted Lagrange map, 34 Riemann-Hilbert problem, 120 Riemann-Hilbert-Birkhoff problem. 121 right equivalent, 64 saturated lattice, 116, 118 second structure connection, 149, 204 semisimple, 10 semiuniversal unfolding, 63, 165 simple F-manifold, 55, 77 small quantum cohomology, 211 smooth divisor, 132 socle field, 249 spectral number, 114, 128, 193, 256 spectral pair, 193 spectrum of a Frobenius manifold, 84, 147 spectrum of a singularity, 172 splitting lemma, 67 stably right equivalent, 67 standard form, 43 strict morphism. 185 symmetries of singularities, 235 Torelli type conjecture, 225, 239 Torelli type result, 226 unfolding, 62 unit field, 14 V-filtration, 112 variance, 256 variation operator, 189 versal Lagrange map, 33, 90 versal unfolding, 63 Virasoro constraints, 252 weight filtration, 183