Solomon
Solomon
On Finite Simple
                      Groups and Their
                         Classification
                                                                                                           by Ron Solomon
         D
                       anny Gorenstein called it the “Thirty      linear, unitary, symplectic, and orthogonal groups
                       Years War,” for the Classification         acting on a finite-dimensional vector space over
                       battles were fought mostly in the          a finite field, whose derived subgroups are all
                       decades 1950–1980, although the            simple, except in a few cases for very small di-
                       dream of a classification of all finite    mensions and fields. The
         simple groups goes back at least to the 1890s.           work of Chevalley, Tits,
         In this brief article, I shall attempt to give some      Steinberg, Suzuki, and
         sense of the mathematical highpoints of the              Ree in the 1950s pro-
         original proof and the ongoing revision project.         vided a systematic de-            . . . the study of
         I shall also give some personal reflections on the
         sociology of the classification effort, and finally
                                                                  scription of finite ana-
                                                                  logues for all of the
                                                                                                     simple groups
         I shall discuss some current and future directions       complex simple Lie              generated amazing
         for research in finite group theory. Many thanks         groups, some real forms
         are due to Jon Alperin, Michael Aschbacher,              and related “twists”. Fi-         insights into the
         George Glauberman, Bill Kantor, Radha Kessar,
         Richard Lyons, and Steve Smith for valuable cri-
                                                                  nally in the period
                                                                  1965–1974, a great gold
                                                                                                   structure of finite
         tiques of this article.                                  rush unearthed twenty-               groups and
             But first, a word about our sponsors: the fi-        one new simple groups
         nite simple groups themselves. Before there even         to supplement the five           uncovered several
         was a mathematical term “group”, Lagrange,               which had been discov-
         Gauss, and their contemporaries were familiar            ered in the 1860s by
                                                                                                        of the most
         with the cyclic groups Zp and the alternating            Mathieu and dubbed              fascinating objects
         groups An. Galois, who gave us the term “group”          “sporadic” by Burnside.
         and the concept of a normal subgroup, was also           Like the elementary par-       in the mathematical
         familiar with the fractional linear groups               ticles of physics, spo-
         PSL(2, p) and their connection to p -division            radic simple groups were
                                                                                                       firmament.
         points on elliptic curves. Jordan described the          often predicted several
         classical linear (or matrix) groups over prime           years before their exis-
         fields, and this was extended to all finite fields       tence was confirmed. For
         by Dickson. This yields the projective special           example, the Monster was predicted in 1973,
                                                                  but not constructed until 1980.
          Ron Solomon is a professor of mathematics at The           A vast literature of theorems, most of which
          Ohio State University, Columbus, OH . His e-mail ad-    were published between 1955 and 1983, com-
          dress is solomon@math.ohio-state.edu.                   bines to yield the following result.
                            The Classification of the Finite Simple Groups                The battle to restrict the structure of H was
                            Let G be a finite simple group. Then G is one of          intricately interwoven during the decade
                            the following:                                            1965–1975 with the quest for new sporadic sim-
                            1. a cyclic group of prime order, Zp;                     ple groups to create a unique mathematical ta-
                            2. an alternating group, An , n ≥ 5 ;                     pestry. This precis will focus only on the battle.
                            3. a classical linear group PSL(n, q), PSU(n, q) ,        This then was one of the principal remaining
                               PSp(2n, q) or PΩε (n, q) ;                             challenges of the Classification problem in 1965:
                            4. an exceptional or twisted group of Lie type            How does one use the hypothesis of simplicity of
                               3 D (q) , E (q) , 2 E (q) , E (q) , E (q) , F (q) ,
                                   4        6         6       7       8     4         G to restrict the structure of the centralizer H ?
                               2 F (2n )0, G (q) , 2 G (3n ) or 2 B (2n ) ;               For example, H could hypothetically have ar-
                                  4          2         2           2
                            5. a sporadic simple group: M11, M12, M22, M23 ,          bitrarily complicated solvable normal subgroups.
                               M24 (the Mathieu groups); J1 , J2 , J3 , J4 (the       But in the actual finite simple groups, solvable
                               Janko groups); Co1, Co2, Co3 (the Conway               normal subgroups of centralizers of involutions
                               groups); HS , Mc ; Suz ( Co1 ‘babies’), Fi22 ,         are of very restricted type. In particular, normal
                                         0                                            subgroups of odd order in centralizers of invo-
                               Fi23 , Fi24   (the Fischer groups); F1 = M (the
                               Monster), F2 , F3 , F5 , He(= F7 ) (Monster ‘ba-       lutions are cyclic and “almost” central. There is
                               bies’); Ru ; Ly ; ON .                                 a suggestive analogy in the theory of Lie algebras:
                                                                                      If L is a finite-dimensional semisimple Lie alge-
                            The Mathematics of the Classification                     bra over C and x is a semisimple element of L
                            The term “Thirty Years War” is apt, inasmuch as           (i.e., ρ(x) is diagonalizable for any matrix rep-
                            the first close approximation to the eventually           resentation ρ of L ), then the centralizer CL (x)
                            successful classification strategy was proposed           is a reductive Lie algebra (i.e., any solvable ideal
                            by Richard Brauer at the International Congress           of CL (x) is central). The achievement of the anal-
                            of Mathematicians in Amsterdam in 1954. His               ogous theorem for finite simple groups (the Bp -
                            idea was:                                                 Theorem) is the longest chapter in the entire clas-
                                                                                      sification and centers around three themes:
                               In a finite nonabelian simple group G ,
                               choose an involution z (an element of order            1. Signalizer Method
                               two) and consider its centralizer
                                                                                      The signalizer method provides the most far-
                               CG (z) = {g ∈ G : gz = zg} . Show that the
                                                                                      reaching answer to the question:
                               isomorphism type of CG (z) determines the
                               possible isomorphism types of G .                         How can one exploit the absence of solvable
                                                                                         normal subgroups in G to bound the struc-
                               During the period 1950–1965, Brauer and oth-
                                                                                         ture of solvable normal subgroups of cen-
                            ers honed methods for solving the class of prob-
                                                                                         tralizers H = CG (z) ?
                            lems: Given a specific H , determine up to iso-
                            morphism all simple groups G having an                       The key initial ideas for the study of “ A -sig-
                            involution z with CG (z) ≡ H . Once Feit and              nalizers” appeared in the work of Thompson,
                            Thompson proved Burnside’s odd order conjec-              while the concept of a signalizer functor is due
                            ture in 1963, Brauer’s strategy gained further            to Gorenstein. The crucial Signalizer Functor
                            credibility: at least one could find an involution        Theorem (see Appendix) gives conditions under
                            in any nonabelian finite simple group!                    which a collection of A -invariant p0-subgroups
                               The subgroup CG (z) is an example of a ( p-)           of G can be glued together into a single proper
                            local subgroup of G , i.e., the normalizer of a non-      p0-subgroup of G . ( X is a p0-group if the prime
                            identity p-subgroup of G for some prime p. (In            p does not divide the order of X.) The wished-
                            this case, p = 2 .) Brauer’s philosophy represented       for conclusion is that this subgroup, Θ(G; A) , is
                            the first version of a type of local-global princi-       a normal subgroup of G , whence Θ(G; A) = 1 .
                            ple that was to determine the shape of the Clas-          This is a lengthy journey, a principal way-station
                            sification proof.                                         of which is the proof that if Θ(G; A) is not nor-
           2. Strong p-Embedding
           Definition. Let G be a finite group. A proper sub-
           group M of G is a strongly p-embedded subgroup
           of G if p divides |M|, but p does not divide
           |M ∩ M g | for any g ∈ G − M .
              This means that in the transitive permuta-
           tion action of G on the coset space M\G, every
           element of order p fixes exactly one point. The
           evolution of this theory passes through the work
           of Frobenius, Zassenhaus, Feit, Brauer, Suzuki,
           and others and, for p = 2 , reaches a very elegant
           conclusion in the Strongly Embedded Theorem
           of Bender. This identifies all simple groups with
           a strongly 2-embedded subgroup and, in partic-
           ular, asserts that no simple group has a strongly
           2-embedded 20 -local subgroup (by a p0-local sub-
           group we mean the normalizer of a nonidentity
           p0-subgroup).
              When p is odd, the story is messier. For ap-
           plications to signalizer theory, the crucial fact is:    In 1982, Danny Gorenstein (above), Ron Solomon, and
           No simple group G of p-rank ≥ 3 has a strongly           Richard Lyons began a “Revision Project” intended to
           p-embedded p0-local subgroup. This has only              produce a “new and complete proof of the Classification.”
           been established after the fact of the Classifica-
           tion. However, in the inductive context, the re-
           quired special case of this result was established       volutions t in G are semisimple (in fact, diago-
           by Aschbacher. Because nontrivial signalizer             nalizable) and a typical such t and its central-
           functors lead to strongly p-embedded p0-local            izer H = CG (t) are:
           subgroups, they in turn do not in general exist.                                                ¥
                                                                                       −Im×m           0
           This is the key to establishing the crucial Bp -The-               t=
           orem whose statement we shall now approach.                                   0         Ir ×r
                                                                                   (           ¥
                                                                                        A 0
           3. Semisimple Elements and Components                             H=                    :
                                                                                        0 B
           Definitions. We call a finite group L quasi-                                                          )
           simple if L = [L, L] and L/Z(L) is simple. A good                           A ∈ GL(m, q), B ∈ GL(r , q)
           example is SL(n, q) for (n, q) 6= (2, 2) or (2, 3).
              We call L a component of H if L is quasi-sim-                    ≡ GL(m, q) × GL(r , q),
           ple and L is a subnormal subgroup of H , i.e.,
                                                                             n = m + r.
           there is a normal series (in the sense of Jordan-
           Holder) from L to H . Distinct components of H           If m > 1 and (m, q) 6= (2, 3) , then SL(m, q) =
           commute and the (commuting) product of all               E(GL(m, q)) is a quasisimple component of H ,
           components of H is denoted E(H) .                        and likewise for SL(r , q) . Thus, except for the
              Components play a dominating role in the              small cases noted, E(H) ≡ SL(m, q) × SL(r , q) =
           centralizers of semisimple elements in classical         [H, H] .
           linear groups (indeed, in all groups of Lie type).          In contrast to this, if G = GL(n, 2m ) , then in-
           For example, if G = GL(n, q) with q odd, then in-        volutions t in G are unipotent matrices and the
                            centralizer of t has a large “unipotent radical”—            With the Bp -Theorem and the p-Component
                            a normal 2-subgroup—and has no components.                Theorems in hand, Brauer’s original strategy can
                                The early work of Feit, Suzuki, and Thompson          now be vindicated and refined to an inductive al-
                            dealt exclusively with groups whose local sub-            gorithm for classifying those finite simple groups
                            groups H were solvable, thus for which E(H) = 1.          that contain a semisimple p-element for some
                            It was Gorenstein and Walter in the mid ’60s who          prime p:
                            first came to grips with the general problem of              1. Choose a semisimple element x of prime
                            simple groups with nonsolvable local subgroups.           order p whose centralizer contains a large com-
                                Definitions. The generalized Fitting subgroup         ponent K , as promised by the p-Component
                            of H is F ∗ (H) = F(H)E(H) where F(H) is the Fit-         Theorem. By induction, K is a known quasi-
                            ting subgroup of H , i.e., the (unique) largest nor-      simple group and by the p-Component Theorem,
                            mal nilpotent subgroup of H. We call a p-element          K is almost all of CG (x) .
                            x of G semisimple if E(CG (x)) 6= 1 . We call a p-           2. Now for each known quasi-simple group K
                            element x of G unipotent if F ∗ (CG (x)) is a p-          and each prime p, classify all finite simple groups
                            group.                                                    G having an element x of order p with CG (x) ap-
                                Caveat: If G is a classical linear group, this no-    proximately equal to K.
                            tion of semisimple roughly corresponds to the                With some refinements (in particular, one
                            classical notion, but there are definite discrep-         must choose p = 2 , if possible), this is the strat-
                            ancies. In a simple classical linear group over a         egy which handles roughly half of the Classifi-
                            field of characteristic p, every nonidentity unipo-       cation proof, but does not handle:
                            tent element (in the classical sense) is unipotent
                            in the above sense.                                       4. Quasi-Unipotent Groups
                                The principal application of the signalizer           Definition. We call G quasi-unipotent if every el-
                            method is to establish a slightly weakened ver-           ement of G of order p is unipotent for all primes
                            sion of the following theorem:                            p such that G has p-rank ≥ 3 .
                            Theorem. If G is simple of p-rank ≥ 3 , then ei-             Thus the other half of the Classification prob-
                            ther some x ∈ G of order p is semisimple or every         lem is the determination of all quasi-unipotent
                            X ∈ G of order p is unipotent.                            groups. In the context of the classification of
                                                                                      finite-dimensional semisimple Lie algebras over
                               At the heart of this analysis is the Bp -Theo-
                                                                                      C, a roughly analogous problem is quickly re-
                            rem. (See Appendix.)
                                                                                      solved by Engel’s Theorem: A finite-dimensional
                               In the semisimple setting a second important
                                                                                      Lie algebra L, all of whose elements are ad-nilpo-
                            answer to the question: How does one exploit the
                                                                                      tent, is itself a nilpotent Lie algebra (hence, in par-
                            simplicity of G ? is provided by the following the-
                                                                                      ticular, L is not semisimple). In the classification
                            orem of Aschbacher (as refined by Foote):
                                                                                      of the finite simple groups, this problem is con-
                               Aschbacher’s Component Theorem: Suppose
                                                                                      siderably thornier. It sits logically at the base of
                            that E(G) is simple and G contains a semisim-
                                                                                      the entire problem, in the sense that any mini-
                            ple involution. Then there is some semisimple
                                                                                      mal simple group is quasi-unipotent.
                            involution x such that CG (x) has a normal sub-
                                                                                         [The classification of the minimal simple
                            group K which is either quasi-simple or iso-
                                                                                      groups was achieved in the monumental Odd
                            morphic to O + (4, q)0 and such that Q = CG (K) is
                                                                                      Order Theorem of Feit and Thompson and the
                            tightly embedded (i.e., |Q ∩ Qg | is odd for all
                                                                                      N -Group Theorem of Thompson.]
                            g ∈ G − NG (Q)) .
                                                                                         There are four principal cases of the general
                               In practice, when K is a known quasi-simple
                                                                                      Quasi-unipotent Problem:
                            group, tight embedding usually implies that Q
                            has 2-rank 1. In the current revision of the Clas-        A. The Odd Order Case, treated by Feit and
                            sification proof, Aschbacher’s theorem is ex-             Thompson.
                            tended to (somewhat different) p-Component                B. The 2-Rank 2 Case, treated by Alperin, Brauer,
                            Theorems for all primes p.                                Gorenstein, Walter and Lyons.
         The remaining two cases, together with many of           The Sociology of the Classification
         the fundamental ideas for their solution, first          in the 1970s
         emerged clearly in Thompson’s fundamental
         work on “ N -groups”.                                           Up to the early 1960s, really
                                                                         nothing of real interest was
         C. The Classical Klinger-Mason Case: G is quasi-                known about general simple
         unipotent of 2-rank ≥ 3 and some 2-local sub-                   groups of finite order. . . .
         group M has p-rank ≤ 2 for every odd prime p.                   Since [1962], finite group the-
         (The “thin” subcase is when the 2-local p-rank                  ory simply is not the same any
         of G is 1.)                                                     more.
         D. The Classical Quasi-Thin Case: G is quasi-                         —Richard Brauer (ICM, 1970)    Richard Brauer
         unipotent of 2-rank ≥ 3 but every 2-local sub-           The Odd Order Theorem of Feit and
         group P has p-rank ≤ 2 for every odd prime p.            Thompson (followed by Thompson’s N -
         (The “thin” subcase is when the 2-local p-rank           group paper) was a singularity in the
         of G is 1.)                                              evolution of finite group theory. An
             The first step in Case D is to establish that        understanding of the dramatic new ideas
         G is generated by two 2-locals, say P1 and P2 ,          and methods introduced in this 255-page
         containing a common Sylow 2-subgroup T of G.             paper became almost indispensible for
         The existence of at least two maximal 2-locals           continued participation in the
         containing a given Sylow 2-subgroup T is guar-           Classification endeavor. Gorenstein wrote
         anteed by the Global C(G; T )-Theorem: If a Sylow        the ‘Reader’s Guide’ in 1968: Finite
         2-subgroup T of G lies in a unique maximal 2-            Groups. He also provided the optimism,
         local P of G , then P is a strongly embedded             the organization and, in 1972, a ‘16-step
         subgroup of G and G is known.                            plan’ for the completion of the
             The Quasi-thin Theorem asserts that in Case          Classification proof. Although obsolete
                                                                                                              Michael Aschbacher
         D , if G does not have a strongly embedded sub-          in several important points within months
         group, then G is a group of Lie type in charac-          of its articulation, Gorenstein’s program
         teristic 2 of Lie rank 2 generated by a pair of (par-    was a critical source of problems and
         abolic) subgroups P1 and P2 or G is on a short           inspiration for the ‘young Turks’ who
         list of exceptions. The strategy is to construct the     attacked the Classification in the ’70s.
         “building”, i.e., the (P1 , P2 ) -coset geometry, and    Yet another critical new feature of the
         then identify this geometry and the associated           ’70s, most notably in the work of
         group. The original proof of the Quasi-thin The-         Timmesfeld and Aschbacher, was the
         orem by Aschbacher and Geoff Mason (only in              fusion of the geometric methods of
         preprint, except for [A1]) uses the Weak Closure         Fischer, Hall, and Shult with the
         Method, introduced by Thompson and extended              architectonic analysis of Thompson,
         by Aschbacher. Current work towards a new                Gorenstein, and Walter.
         proof uses the Amalgam Method, introduced by                The pace of the Classification in the
         Goldschmidt. These methods afford the final              ’70s was exhilarating. Not a single lead-
         deep answer to the question:                             ing group theorist besides Gorenstein be-   John Thompson
                                                                  lieved in 1972 that the Classification
            How does one exploit the simplicity of G to           would be completed this century. By         Numerous
            bound the p-local structure of a (quasi-              1976, almost everyone believed that the     mathematicians have
            unipotent) simple group G ?                           Classification problem was “busted”. The    been directly or
                                                                  principal reason was Michael Asch-          indirectly involved in
         This completes a very brief overview of the strat-       bacher’s lightning assaults on the B-Con-   this research, many of
         egy for the Classification proof.                        jecture, the Thin Group Problem, and the    their names appear in
                                                                  Strongly p-embedded 2-local problem.        gray along the tops of
                                                                  Also, in 1976 Timmesfeld announced a        these pages.
         working of the Odd Order Theorem and the                chbacher has taken a first major step towards
         Abelian and Dihedral Sylow 2-Subgroup Theo-             codifying this knowledge in [A2]. There is also a
         rems in the late ’60s. In so doing, he enriched all     book on the sporadic groups in preparation by
         of group theory with fundamental new con-               Griess.)
         cepts—e.g., F ∗ (G) —and new theorems. Later,               The GLS work itself will appear in a series of
         several others, notably Glauberman,                                approximately twelve volumes to be
         Peterfalvl, and Enguehard undertook                                published by the American Mathemat-
         various “revision” projects. This effort     @?
                                                      @L
         vision of both the Odd Order Theo-                                     Part I consists of: overview and out-
                                                      5?e?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@X?
                                                      Y?e?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@,?                  ?3@@@@@@@@T@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?@@g
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@(Y?                 ?V'@@@@@@S@U@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?@@g
                                                                                                                              V4@@@@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?@@e?W2@
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@H
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@W.?                        @@(R'>(?'@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?@@e?7@@
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@U?                        @0Y?S@H?N@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?@@e?@@@
                                                                                                                                      7@L??@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
                                                      @@@5?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@1?
                                                      @@0Y?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@W2@                          @@@?e?@@@@@@@@@@@@@@@@@@@@@@@@@@@@?@@e?3@@
                                                           ?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?/X?                     3@f?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?V'@
                                                           ?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@W@@@@?N1?heW26XgV'1?e?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@fN@
                                                      @@6X?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@8W@@L?@?h?W.MB)K?f?V'?@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@f?@
                                                      @@@1?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@T&@@@?he?.Y??@@6X?g@@?3@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@V@@@X                    I')Kg@@?S@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
                                                                                                                                ?N@@@6X?e@@W&R4@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@1
                                                      @@@5?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@f?W26K?eO@K?f@?@@1?e@@@@eW@@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
                                                      @@0Y?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@6KO&@@@@@@@@@@e?J@?3@@Le@@@@e7@@@@@@@@@@@@@@@@@@@@@@@@@@@e?3@@
                                                           ?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@??@@@@@@@@@@@@@@@@@@@@@e?7@?N@@1e@@@@e@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?V4@
                                                           ?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?f?3@??@@@e@@@@L?3@@@@@@@@@@@@@@@@@@@@@@@@@@@g
                                                      @6K??@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@V4@@@@@@@@@@6K??V'?g3@@@1?V'@@@@@@@@@@@@@@@@@@@@@@@@@@f?J
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?e@@@@@@@@@@@@heV@@@@?e@@@@@@@@@@@@@@@@@@@@@@@@@@fO&
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@(MW@@@@@Le@@@@@@@@@@@@L?g?@@@@@@?O&@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
         rem and the identification of the split                            line, general group theory, and prop-
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@YO&Y@@@@@f@@@@@@@@@@)Xg?N@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@X@@@@Hf@@@@@?@@@@@1g?J@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@??S@@@@@?f@@@@@@@@W@@@g?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
                                                           ?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@V4@?W&@@@@@?f?W@@@@@@(Mh?N@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?3@@
                                                           ?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@L??W&@@@@@5?f?'@?@@@0Y?h?J@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?V'@
                                                           ?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@)X?*@@@@@(Y?f?V'@@@h?@e?7Y@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@fN@
                                                      @@6X?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@(e@@@@@HhV'@@@@(?f?@e?@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@fJ@
                                                      @@@1?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@(Y?@@@@@@?e?@f?V4@@0Y?he?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?W&@
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@Y??N@@@@5?e?@?@K?                         @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?7@@
                                                      @@@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?J@@@@g?@@@@?f?'@?e?W2(e@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?@@@
                                                      @@@5?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@H?7@@@@)KeO@X@@?g?V'?e?&0Ye@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?3@@
                                                      4@0Y?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@??@@@@@@@@@@@@@@@6K                  @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@e?N@@
                                                           ?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@5?J@@@@@@@@@@@@@?@@@@?he?W2@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@f@@
                                                           ?@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@YO&@@@@@@@@@@@@@?3@@@?e@?f?@?7@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@5g
                                                           ?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@(Me?V'@@@@?h?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?g
                                                      @@6X?@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@H?f?@@@H?f?)X?e?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@1g
                                                      @@@1?@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@T&K??O&@@@W.?@eJ@(?@6X@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@5e?W2@
                                                      @@@@?@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?V@@@@@@@@@@@Ue?O&(YJ@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@HeW&@@
                                                      @@@5?@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@1e@@@HW&@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?e7@@@
                                                      @@(Y?@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@XI@e@@@W&@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?e3@@@
                                                      @(Y??@@?@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@fI4@@@@f@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@?eN@@@
         schmidt introduced the Amalgam                                      finite groups of Lie type and will enu-
                                                       @@)X                             O@K?
                                                       @@@1eW2@@6XeW2@@6Xe?W2@@@e?@@@@@eW2@@6Xe?W2@@@e?@@@@@e?@@@@@e?W26K?eW2@@6Xhe?W26K?e?W26X?hfO&
                                                       @@@@?O&@@@@1e7@@@@1e?7@@@@L?J@@@@@L?7@@@@1e?7@@@@L?J@@@@@eJ@@@@@L?W&@@@@e7@@@@1eW2@@6XeW&@@@@eW&@@)Xe?@@@6Xe?@@@
                                                       @@@@@@@@@@@@L?@@@@@@L??@@@@@)?&@@@@@)X@@@@@@e?@@@@@)?&@@@@@?O&@@@@@)?&@@@@@e@@@@@@?O&@@@@1e7@@@@@?O&@@@@1eJ@@@@1eJ@@@
         Method with the principal intention of                              merate the assumed results on spo-
         “revising” the weak closure arguments             Sir W. R.         radic groups. The overview and outline
         at the heart of Thompson’s N -group               Hamilton          volume has been sent to the publisher.
         paper and of the thin and quasi-thin            observed in         The remainder of Part I should be com-
         group work of Aschbacher and Mason.            1856 that the        pleted within the next year.
         This program has had major successes,           icosahedral             Part II presents the fundamental
         notably Stellmacher’s revisions of the       group (left) may       “uniqueness theorems” on which the
         core of Thompson’s N -group paper                be defined         Classification rests: the Suzuki-
         and Aschbacher’s thin group paper. It          abstractly as        Bender Strongly Embedded Theorem
         too has had a fruitful influence on the          the group          (with extensions), the Strongly p-em-
         related fields of finite geometry and          generated by         bedded Theorem (work of Gernot
         geometric group theory.                             two             Stroth), the Global C(G; T )-Theorem
            In 1982, Danny Gorenstein launched        substitutions of       (Joint work with Richard Foote) and tje
         a “revision project” in which he was          orders 2 and 3,       p-Component Uniqueness Theorems.
         joined by Richard Lyons and myself.            respectively,        The main mathematical body of Part
         This project is intended to complement       whose product          II is complete, although much remains
         the work of the other revision efforts         is of order 5.       to be done in terms of preparation of
         to yield a new and complete proof of                                preliminary sections.
         the Classification. Here is a brief dis-                                Part III presents the proof of the
         cussion of the status of this (GLS) project and its     “generic case”. This constitutes the classification
         interfaces with the other revision efforts.             (subject to the inductive assumption that all
            The work of GLS rests on a foundation of             proper simple sections are isomorphic to known
         background results. In addition to the contents         simple groups) of most of the finite simple
         of standard textbooks and monographs, these             groups, including An for n ≥ 13 and the groups
         consist principally of:                                 of Lie type of rank ≥ 4 (except for a few defined
            1. The Odd Order Theorem and the identifi-           over F2 ). Operationally, the generic case treats
         cation of the split BN -pairs of rank 1. (As dis-       those local configurations to which, for some
         cussed above, these theorems have already been          prime p, the Signalizer Method can be effec-
         subjected to extensive review and revision.), and       tively applied to verify the Bp -Property for a
            2. The existence, uniqueness, Schur multipli-        large number of semisimple p-elements of G .
         ers, and other basic properties of the twenty-six       More than half of the mathematical body of Part
         sporadic simple groups. (These properties are           III is complete.
         stated without proof in [GL] and in the Atlas               The “special” (nongeneric) part of the proof
         [Co]. Many have published proofs in journals. As-       divides into “odd” and “even” cases, according