HOW MANY ESSENTIALY DIFFERENT FUNCTION THEORIES EXIST? VLADIMIR V. KISIL Institute of Mathematics, Economics and Mechanics, Odessa State University ul. Petra Velikogo, 2, Odessa-57, 270057, UKRAINE Abstract. The question in the title is ambiguous. At least the understand- ing of words essentially different and function theory should be clarified. We discuss approaches to do that. 1. Introduction ‘The classic heritage of complex analysis is contested between several com- plex variables theory and hypercomplex analysis. The first one was founded Jong ago by Cauchy and Weierstrass themselves and sometime thought to be the only crown-prince. The hypercomplex analysis is not a single theory but a family of related constructions discovered quite recently [3, 5] (and rediscovered up to now) under hypercomplex framework Such a variety of theories puts the question on their classification. One could dream about. a Mendeleey-like periodic table for hypercomplex anal- ysis, which clearly explains properties of different theories, relationship be- tween them and indicates how many blank cells are waiting for us. More- over, because hypercomplex analysis is the recognized background for clas- sic and quantum theories like the Maxwell and Dirac equations, such a table could play the role of the Mendeleev table for elementary particles and fields. ‘We will return to this metaphor and find it is not very superficial ‘To make a step in the desired direction we should specify the notion of function theory and define the concept of essential difference. Probably many people agree that the core of complex analysis consist of ns Dietrich ea. (eds). Cliford Algebras and Their Application in Mathematical Physics, 75-188 (© 1998 Kher Academie Publishers.176 VLADIMIR V. KISIL 1. The Cauchy-Riemann equation and complex derivative 2; 2. The Cauchy theorem; 3. The Cauchy integral formula; 4. The Plemeli-Sokhotski formula; 5. The Taylor and Laurent series. Any development of several complex variables theory or hypercomplex anal- ysis is beginning from analogies to these notions and results. Thus we adopt the following Definition 1.1 A function theory is a collection of notions and results, which includes at least analogies of 1-5. Of course the definition is more philosophical than mathematical. For ex- ample, the understanding of an analogy and especially the right analogy usually generates many disputes. Again as a first approximation we propose the following, Definition 1.2 Two function theories is said to be similar if there is a cor- respondence between their objects such that analogies of 1-5 in one theory follow from their counterparts in another theory. Two function theories are essentially different if they are not similar. Unspecified “correspondence” should probably be a linear map or some- thing else and we will look for its meaning soon. But it is already clear that the similarity is an equivalence relation and we are looking for quotient sets with respect to it. In the rest of the paper we will look for more rigor understanding of it. 2. Abstract Nonsense about Function Theory In this Section we repeat shortly the scheme of development of Clifford analysis as it could be found in (3, 5]. We examine different options arisi on this way and demonstrate that some differences are only apparent not essential. 2.1, FACTORIZATIONS OF THE LAPLACIAN ‘We would like to see how the contents of 1-5 could be realized in a function theory. We are interested in function theories defined in R¢. The Cauchy theorem and integral formula clearly indicates that the behavior of func- tions inside a domain should be governed by their values on the boundary.HOW MANY ESSENTIALY DIFFE! T 17 Such a property is particularly possessed by solutions to the second order elliptic differential operator P 4 4 P(z,0x) = > ais(x)0:9; + J bi(2)0; + e(z) i with some special properties. Of course, the principal example is the Lapla- cian ‘ ae A=>o oF (a) 1. Choice of different operators (for example, the laplacian or the Helm- holtz operator) is the first option which brings the variety in the family of hypercomplex analysis. The next step is called linearization, Namely we are looking for two first order differential operators D and D! such that DD! = P(z,d-). ‘The Dirac motivation to do that is to “look for an equation linear in in time derivative 2, because the Schrédinger equation is”. From the function theory point of view the Cauchy-Riemann operator should be linear also. But the most important gain of the step is an introduction of the Clifford algebra. For example, to factorize the Laplacian (1) we put (2) where €; are the Clifford algebra generators: ee) Fee = By, STIS. (3) 2. Different linea) trum of theories. ions of a second order operator multiply the spec- Mathematicians and physicists are looking up to now new factorization even for the Laplacian. The essential uniqueness of such factorization was already felt by Dirac himself but it was never put as a theorem. So the idea, of the genuine factorization becomes the philosophers’ stone of our times. After one made a choice 1 and 2 the following turns to be a routi The equation D'f(z) =0,178. VLADIMIR V. KISIL plays the role of the Cauchy-Riemann equation. Having a fundamental so- lution F(z) to the operator P(z,d-) the Cauchy integral kernel defined by Elz) = D'F(2) with the property DE(2) = 6(2). Then the Stocks theorem implies the Cauchy theorem and Cauchy integral formula. A decomposition of the Cauchy kernel of the form Czy) = SO Valz)Waly), where Va(2) are some polynomials, yields via integration over the ball the ‘Taylor and Laurent series. In such a way the program-minimum 1-5 could be accomplished. Thus all possibilities to alter function theory concentrated in 1 and 2. Possible universal algebras arising from such an approach were investigated by F. Sommen [17]. In spite of the apparent wide selection, for operator D and D’ with constant coefficients it, was found “nothing dramatically new”: Of course one can study all these algebras and prove theorems or work out lots of examples and representations of universal algebras. But in the constant coefficient case the most. important factorization seems to remain the relation A = S23, ie., the one leading to the definition of the Clifford algebra. We present an example that there is no dramatical news not only on the level of universal algebras but also for function theory (for the constant coefficient case). We will return to non constant case in Section 3. EXAMPLE OF CONNECTION We give a short example of similar theories with explicit connection be- tween them. The full account could be found in [8], another example was considered in [14]. Due to physical application we will consider equation of “8 Oe ogg toh () where e; are generators (3) of the Clifford algebra and M = My is an operator of multiplication from the right-hand side by the Clifford number 2. Equation (4) is known in quantum mechanics as the Dimac equation for 4 particle with a non-zero rest mass (1, §20}, (2, §6.3] and [12]. We willHOW MANY ESSENTIALY DIFFERENT 179 specialize our results for the case M = Ma, especially for the simplest (but still important!) case A € R. ‘Theorem 2.1 The function f(y) is a solution to the equation Of gt 8 s—= a+ M ip Dest of if and only if the function aly) = Mae“ ™™ fy) is a solution to the equation where My and My are bounded operators commuting with e; Corollary 2.2 The function f(y) is a solution to the equation (4) if and only if the function e®™ f(y) is a solution to the generalized Cauchy- Riemann equation (2) In the case M = My we have &®™> f(y) = f(y)e® and if € R then 20M Fly) = Fly)e® = f(y) In this Subsection we construct a function theory (in the sense of 1 5) for M-solutions of the generalized Cauchy-Riemann operator based on Clifford analysis and Corollary 2.2. ‘The set of solutions to (2) and (4) in a nice domain @ will be denoted by MQ) = Mo(M) and Myy(M) correspondingly. In the case M = My we use the notation My(2) = May,(®) also. We suppose that all functions from 97\() are continuous in the closure of . Let — PCH) ae ODA fy a Ely-2) (5) be the Cauchy kernel [5, p. 146] and do = S7(-1ejdaaA...A [dai] A... A dtm = be the differential form of the “oriented surface element” (5, p. 144]. Then for any f(z) € (2) we have the Cauchy integral formula [5, p. 147] S(z), rEeQ By 2)do, st) = {10 288. sa.180 VLADIMIR V. KISIL Theorem 2.3 (Cauchy’s Theorem) Let f(y) € Mxy(M). Then [ doy e¥™ f(y) = 0. loa Particularly, for f(y) € 24(2) we have J aovstane™ ha and [toe Sea) = 9, ba ifteR. Theorem 2.4 (Cauchy’s Integral Formula) Let f(y) € May(Q). Then * nM yy) = { Sah 22 em [ ry-advemsn= {10 288.) Particularly, for f(y) € 9,(Q) we have E(y ~ (2o-w)r _ { f(z), rea ff Blu~ ado, paneer = { 1! 268 and L Ely — 2)el=0-) da, so={ S(z), reo ba y 0, go" ifA ER. It is hard to expect that formula (6) may be rewritten as wy _f fle), 2€2 ff Pu-ardesw = {10> 258 with a simple function E'(y — 2). Because an application of the bounded operator ¥™ does not destroy uniform convergency of functions we obtain (cf. (5, Chap. II, § 0.2.2, The- orem 2)) Theorem 2.5 (Weierstrass’ Theorem) Let {f,}sen be a sequence in Myr(), which converges uniformly to f on each compact subset K € Q. Then 1. f € Mu (Q).HOW MANY ESSENTIALY DIFFERENT 181 2, For each multi-index 8 = (90,--.,8m) €N"*", the sequence {3° fi}ken converges uniformly on each compact subset K € Q to "J. ‘Theorem 2.6 (Mean Value Theorem) Let f € Myy(). Then for all + €Qand R>0 such that the ball B(z, R) € 2 om (1+ )T(2H) fe)= Sessariie iy NSO 3. Hypercomplex Analysis and Group Representations — To- wards a Classification To construct a classification of non-equivalent objects one could use their groups of symmetries. Classical example is Poincaré's proof of bi- holomorphic non-equivalence of the unit ball and polydisk via comparison their groups of bi-holomorphic automorphisms. To employ this approach wwe need a construction of hypercomplex analysis from its symmetry group. The following scheme is firstly presented here (up to the author knowledge) and has its roots in (7, 9, 10}. Let G be a group which acts via transformation of a closed domain ©. Moreover, let G : 2 -+ 89 and G act on © and 0 transitively. Let. us fix a point zo € @ and let H CG be a stationary subgroup of point zp. ‘Then domain @ is naturally identified with the homogeneous space Gi/H. Till the moment we do not request anything untypical Now let there exist a H-invariant measure dj on A. We consider the Hilbert space Lo(22, dj). Then geometrical transformations of 02 give us the representation x of G in Lo(0,dy). Let folx) = 1 and Fx(09,du) be the closed liner subspace of L2(J, dy.) with the properties: 1. fo € Fa(62, dys); 2. Fx(00, dy) is G-invariant; 3. F(A, dj) is G-irreducible. e standard wavelet transform W is defined by W : Px(02, dpt) > Lo(G) : f2) 4 Fla) = (Fle). *(g)fol2))a(00.a) Due to the property [*(h)fol(z) = fol2), h € H and identification Q ~ G/H it could be translated to the embedding: W : F,(0, dp) + L2(Q) : (2) 4 Fy) = UF). 79) fol2))ratonayy> (1) where y € for some h € H. The imbedding (1) is an abstract analog of the Cauchy integral formula. Let functions V be the special functions182 VLADIMIR V. KISIL generated by the representation of H. Then the decomposition of fo(y) by Va gives us the Taylor series. ‘The scheme is inspired by the following interpretation of complex anal- ysis. Example 3.1 Let the domain 2 be the unit disk D, AD = 8. We select the group SL(2,R) ~ SU(1,1) acting on D via the fractional-linear trans- formation: ab) ath cd) ** 4d We fix 29 = 0. Then its stationary group is U(1) of rotations of D. Then the Lebesgue measure on § is U(1)-invariant. We obtain D ~ SL(2,R)/U(1). The subspace of La(S, dt) satisfying to 1-3 is the Hardy space. The wavelets transform (1) give exactly the Cauchy formula. The proper functions of U(1) are exactly 2", which provide the basis for the Taylor series. The Riemann mapping theorem allows to apply the scheme to any connected, simply- connected domain. The conformal group of the Mdbius transformations plays the same role in Clifford analysis. One usually says that the conformal group in R", n > 2is not so rich as the conformal group in R?. Nevertheless, the conformal covariance has many applications in Clifford analysis [4, 15]. Notably, groups of conformal mappings of open unit balls B® C R" onto itself are similar for all n and as sets can be parametrized by the product of B® itself and the group of isometries of its boundary S"~!. ‘Theorem 3.2 [11] Let a € B", 6 € P', then the Mébius transformations of the form boo 1 na b ~ba ton= (Sah )(L)= (ude al): constitute the group B, of conformal mappings of the open unit ball B” onto itself. B, acts on B" transitively. Transformations of the form $40.) consti tute a subgroup isomorphic to O(n). The homogeneous space By /O(n) is isomorphic as a set to B®. Moreover: 1. Gq) = 1 identically on B” (chy = $(01)). 2 ba,1)(0) = & Ha, (2) = 0. Obviously, conformal mappings preserve the space of null solutions to the Laplace operator (1) and null solutions the Dirac operator (2). TheHOW MANY ESSENTIALY DIFFERENT 183 group B, is sufficient for construction of the Poisson and the Cauchy inte- gral representation of harmonic functions and Szegé and Bergman projec- tions in Clifford analysis by the formula (9] f where ry is an irreducible unitary square integrable representation of @ group G, f(z) is an arbitrary non-zero function, and c is a constant. K(2,) Ira fl(z) lr FI) dg, (2) ReMARK 3.3 It is known that many results in several variables theory could be obtained or even explained via hypercomplex analysis (13). One could see roots of this phenomenon in relationships between symmetry groups of two theories: the symmetry group of hypercomplex analysis is wider. Returning to our metaphor on the Mendeleev table we would like recall that it began as linear ordering with respect to atomic masses but have receive an explanation only via representation theory for the rotation group. Our consideration provide a ground for the following Conjecture 3.4 Most probably there is the only constant coefficient func- tion theory or at most there are two of them. |. Acknowledgments and Apologies It is my pleasure to express my gratitude to the R.Delanghe, J. Cnops, K. Guerlebeck, V.V. Kravchenko, I. Mitelman, J. Ryan, M. Shapiro, F. Sommen, W. Sprétig, N. Vasilevski who share with me their ideas on hy- percomplex analysis. It was especially stimulating for this paper author's stay at Universiteit Gent, Vakgroep Wiskundige Analyse (Belgium) under INTAS grant 93-0322 ‘The bibliography to such a paper should be definitely more complete and representative. Unfortunately, | mentioned only a few papers deserving it. References 1. Berestetskii V. B., Lifshitz B. M., and Pitaevski L. P. (1982) Quantum Electrody- namics, volume 4 of L. D. Landay and E. M. Lifshitz, Course of Theoretical Physics. Pergamon Press, Oxford, second edition 2. Bogoliubow N. N. and Shirkov D. V. (1980) Introduction to the Theory of Quantized Fields, John Willey & Sons, Inc., New York, third edition 3. Bracke F., Delanghe R. and Sommen F. (1982) Clifford Analysis, volume 76 of Research Notes in Mathematics, Pitman Advanced Publishing Program, Boston.184 10. a. 12, 13. M4. 15. 16. v. VLADIMIR V. KISIL Chops J. (1994) Hurwitz Pairs and Applications of Mobius Transformations, Habil- itation dissertation, Universiteit Gent, Faculteit van de Wetenschappen. Delanghe R. Sommen F., and Souéek V. (1992) Clifford Algebra and Spinor- Valued Functions, Kluwer Academic Publishers, Dordrecht. Keller J. (1993) The geometric content of the electron theory. part i. Adv. in Appl Clifford Algebras 8(2), 147-200. Kisil V. V. (1994) Relative convolutions. I. Properties and applications., Reporte Interno # 162, Departamento de Matematicas, CINVESTAV del LP.N., Mexico City. e-print archive fuct~an/9410001, to appear in Advances in Mathamatics. Kisil V. V. (1995) Connection between diflerent function theories in clifford analysis, Adv. in Appl. Clifford Algebras 5(1), 63-74, e-print archive funct-an/9501002 Kisil V. V. (1995) Construction of integral representations in spaces of analytical functions, Dokl. Akad. Nauk SSSR, 18, (To appear. Kisil V. V. (1995) Integral representation and coherent states, Bull. Soe. Math. Belg Sér. A, 18, (To appear) Kisil V. V. (1996) Mobius transformations and monogenic functional calculus, Electr. Research Announcements of AMS 2(1), (To appear), Kravchenko V. V. (1995) On biquaternionic bag model, Zeitschrift fir Anal. und ihre Anwend. 14(1), (To appear). Mitelman I. and Shapiro M. (1995) Differentiation of the Martinelli-Bochner integral ‘and the notion of hyperdifferentiability, Math. Nachr. 172, 211-238, Ryan J. (1993) Intrinsic Dirac operator in Cn, Advances in Mathematics, 43, (To appear). Ryan J. (1995) Some application of conformal covariance in Clifford analysis, in J. Ryan, Clifford Algebras in Analysis and Related Topics, 128-155, CRC Press, Boca Raton, Shapiro M. and Vasilevski N, (1994) Quaternionic ¥-hypetholomorphie functions, singular integral operators and boundary value problems. I. ¢-Hypetholomorphic function theory, Compler Variables Theory Appl. 27. Sommen F. (1995) Universal algebras coming from factorization, 10, Preprint