Remembering Leon Ehrenpreis (1930–2010)
Total Page：16
File Type：pdf, Size：1020Kb
Remembering Leon Ehrenpreis (1930–2010) Daniele C. Struppa Leon Ehrenpreis: A Note on His Mathemat- ical Work Leon Ehrenpreis passed away on August 16, 2010, at the age of eighty, after a life enriched by his passions for mathematics, music, running, and the scriptures. I first met Leon in 1978, when I was a doctoral student under Carlos Berenstein (himself a student of Leon’s), at a time when Carlos had invited him to lecture at the complex analysis seminar at the University of Maryland. I had just begun to attempt my first (of many) readings of his work on the fundamental principle [11] and I was in awe of his presence. And yet, I discovered a gentle human being, with a passion for mathematics and a genuine desire to help the newcomer (me, in that case) to understand its mysteries. Photograph by François Rouvière. What follows in this section is a quick overview Leon explaining his AU-spaces at the 2008 of Ehrenpreis’s work, mostly biased by my own conference held in Stockholm in honor of Jan interest and work on the fundamental principle. Boman’s seventy-fifth birthday. It is of course impossible to pay full tribute to the complexity of Ehrenpreis’s work in such a short note. The subsequent sections are short skill in his willingness to tackle bold generaliza- remembrances, both mathematical and personal, tions, as he extends distribution theory to the of Leon’s life, from some of his friends and case in which Euclidean spaces are replaced by collaborators. locally compact Hausdorff spaces denumerable at Leon Ehrenpreis’s mathematical career started infinity and the derivations ∂/∂x are replaced by with his 1953 dissertation [4], which he wrote while i a countable family of local, closed operators. doing his doctoral work at Columbia University He then plunged immediately into the topics under the guidance of Chevalley. The dissertation, that would become the main drivers for his funda- whose main results later appeared in [8], already mental principle. In an impressive series of papers shows some of Ehrenpreis’s most characteristic from 1954 to 1960, Ehrenpreis addressed the prob- Daniele C. Struppa is chancellor at Chapman University, lem of division in a variety of contexts. Specifically, Orange, CA. His email address is [email protected]. he investigated in [5] the issue of surjectivity of 674 Notices of the AMS Volume 58, Number 5 system P1(D)f = · · · = Pr (D)f = 0. The funda- mental principle roughly states that it is possible to obtain a topological isomorphism between the space EP and the dual of a space of holomorphic functions satisfying suitable growth conditions on the union of certain varieties Vk, where {Vk} is a finite sequence of algebraic subvarieties of V = {z : P1(z) = · · · = Pr (z) = 0}. The proofof the theorem is quite complicated, and even decades after Ehrenpreis’s original announcement, it is still the subject of analysis and reinterpretations (see, for example, the chapters dedicated to the fun- damental principle in [2] and in the third edition of Hörmander’s monograph on several complex variables [25]). The crux of the proof, however, Photo by Fabrizio Colombo. is the understanding of the nature of the growth Leon with his student Carlos Berenstein, Carlos’s student Daniele Struppa, and conditions that must be satisfied by holomor- Daniele’s student Irene Sabadini at the 2004 phic functions on the variety V for the duality conference held in Fairfax in honor of Carlos’s to take place. Such growth conditions must be sixtieth birthday. imposed not just on the functions, but also on suitable derivatives that somehow incorporate the multiplicities related to the polynomials Pj . Maybe the most important consequence of this partial differential operators with constant coef- result is the representation theorem, which states ′ ficients in the space DF of distributions of finite that every function f in EP can be represented as orderandinthe space E′ of infinitely differentiable t functions. These were the heroic years of the appli- f (x) = Q (x) exp(iz, x) dµ (z), Z k k cation of the theory of distributions to the study of kX=1 Vk the solvability of linear constant coefficients par- where the Q s are polynomials and µ s are tial differential equations, and Ehrenpreis’s results k k bounded measures supported by V . joined those of Hörmander[23] and Malgrange[27] k It is important to point out that the fundamental to help develop the modern form of this theory. principle actually extends to rectangular systems In [6] Ehrenpreis tackled the case of some special of differential equations and that it holds for convolution equations and extended the results a very large class of spaces (what Ehrenpreis of L. Schwartz on exponential solutions to such called analytically uniform spaces), a category that equations [30], an analysis that he continued later includes Schwartz’s distributions, Beurling spaces, on in [7], where he considered a larger class of holomorphic functions, etc. (when the objects spaces. in these spaces are generalized functions, the All of this culminated with his striking an- representation theorem stated above has to be nouncement [9] in 1960 of what became known suitably interpreted, of course). as the fundamental principle of Ehrenpreis- One of the common themes in most of Ehren- Palamodov (this result was in fact also preis’s work in those years was a sort of philo- independently and concurrently discovered sophical principle that held that one could extend by Palamodov; see [28] and the Russian references most structural results that hold for holomorphic therein) and whose complete proof appeared in functions to spaces of solutions of systems of his most important monograph [11]. linear constant coefficient differential equations The fundamental principle can be seen as a satisfying suitable conditions. far-reaching generalization of the well-known rep- Maybe the most striking such result was his resentation theorem for solutions of ordinary, beautiful and elegant proof of the classical Har- constant coefficients linear differential equations, togs’ theorem on the removability of compact which we teach in any introductory class on singularities for holomorphic functions. By es- differential equations. sentially using the arguments he had developed In order to state it (at least in a particular case), in the course of his study of division problems, let us consider the space E of infinitely differen- Ehrenpreis demonstrated [10] that Hartogs’ phe- n tiable functions on R , and let P1(D), . , Pr (D) be nomenon is not a peculiarity of holomorphic r differential operators with constant coefficients, functions in several variables but in fact holds for whose symbols are the polynomials P1, ··· ,Pn. more general spaces of functions that are solu- Now let EP denote the space of infinitely dif- tions of systems of differential equations (or even ferentiable functions that are solutions of the convolution equations), satisfying some specific May 2011 Notices of the AMS 675 algebraic properties. The paper is beautiful and fifty years. We began as teacher-student when I elegant and shows Ehrenpreis’s talent for simplic- was still doing my undergraduate degree. I heard ity and anticipates the kind of algebraic treatment about a famous, brilliant young mathematician of systems of differential equations that became who was going to give a course entitled Mathe- so important in years to come. matics and the Talmud and I decided to register The interest of Ehrenpreis in removability of for it. After the semester I asked him whether he singularity phenomena is also apparent in another would be willing to write a letter supporting my beautiful series of papers ([12], [13], [14], [15]) application to graduate school, and he told me dealing with elegant and surprising variations on that I would be better off asking a person more the edge-of-the-wedge theorem. familiar with my mathematical abilities. After that I am obviously unable to touch upon all the I did not have much personal contact with Leon important work of Ehrenpreis, but it would be until after I finished my graduate work and our impossible not to mention his interest in the paths crossed again. I was working on problems Radon transform, which absorbed much of his related to compact Riemann surfaces, and we mathematical efforts in the last ten or so years. would meet now and then. Leon was also very His work on this topic culminated in his second much interested in this subject (Ehrenpreis con- monograph [16], in which Ehrenpreis introduces a jecture, Schottky problem) and always expressed generalized definition of the Radon transform in an interest in what I was doing. We even wrote terms of very general geometric objects defined two papers together ([20], [17]), but this is not the on a manifold (what he calls spreads). Just like measure of the influence he had on me. In fact, his previous monograph [11], this more recent my book with Irwin Kra, Theta Constants, Riemann work brings a new and invigorating perspective Surfaces and the Modular Group, [21] grew out of to his subject matter. And as in the case of a conversation with Leon. A visit to New York [11], this work will provide mathematicians with would never be complete without meeting Leon. In ideas and challenges that should keep them busy fact I arrived in New York on August 15 planning for a long time to come. As Ehrenpreis’s former to meet with Leon in the middle of the week to student Carlos Berenstein states in his review [1] discuss my current work, but he passed away on of [16] “[this is] a book that is worth studying, Monday, August 16.