Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

This Week's Finds in Mathematical Physics (Week 254)

  1. Jul 14, 2007 #1
    Also available as http://math.ucr.edu/home/baez/week254.html

    July 13, 2007
    This Week's Finds in Mathematical Physics (Week 254)
    John Baez

    This week I'd like to talk about exceptional Lie algebras and the
    Standard Model, Witten's new paper on the Monster group and black
    holes in 3d gravity, and Connes and Marcolli's new book! Then
    I want to continue the Tale of Groupoidification.

    However, I don't have the energy to do this all now. And even
    if I did, you wouldn't have the energy to read it.

    So, I'll just point you towards Connes and Marcolli's new book,
    which you can download for free:

    1) Alain Connes and Mathilde Marcolli, Noncommutative Geometry,
    Quantum Fields and Motives, available at

    I hope to discuss it sometime, especially since it tackles a
    question I've been mulling lately: is there a good "explanation"
    for the Standard Model of particle physics?

    For now, I'll start by discussing Witten's latest paper:

    2) Edward Witten, Three-dimensional gravity revisited, available
    as arXiv:0706.3359.

    This is a bold piece of work, which seeks to relate the entropy of
    black holes in 3d quantum gravity to representations of the
    Monster group - the largest sporadic finite simple group, with
    about 10^{54} elements.

    If the main idea is right, this gives a whole new view of
    "Monstrous Moonshine" - the bizarre connection between the Monster
    and fundamental concepts in complex analysis like the j-function.
    (See "week66" for a quick intro to Monstrous Moonshine.)

    As the title hints, Witten had already tackled quantum gravity in
    3 spacetime dimensions. In this earlier work, he argued it was an
    exactly soluble problem: a topological field theory called
    Chern-Simons theory. However, this theory is really an
    *extension* of gravity to the case of "degenerate" metrics:
    roughly speaking, geometries of spacetime where certain regions
    get squashed down to zero size. Degenerate metrics are weird.
    So, what happens if we try to quantize 3d gravity while insisting
    that the metric be nondegenerate?

    It's hard to say. So, Witten takes a few clues and cleverly fits
    them together to make a surprising guess. He considers 3d general
    relativity with negative cosmological constant. This has 3d
    anti-DeSitter space as a solution. Anti-DeSitter space has a
    "boundary at infinity": a 2d cylinder with a conformal structure.
    The "AdS-CFT" idea, also known as "holography", suggests that in
    this sort of situation, 3d quantum gravity should be completely
    described by a field theory living on this boundary at infinity -
    a field theory theory with all conformal transformations as

    Which conformal field theory should correspond to 3d quantum
    gravity with negative cosmological constant? It depends on the
    value of the cosmological constant! Some topological arguments
    suggest that the Chern-Simons description of 3d quantum gravity
    is only gauge-invariant when the cosmological constant Lambda
    takes certain special values, namely

    Lambda = -1/(16 k)^2

    where k is an integer, known as the "level" in Chern-Simons theory.

    By the way: I'm working in Planck units here, and I'm assuming
    our Chern-Simons theory is left-right symmetric, just to keep
    things simple. I may also be making some small numerical errors.

    This quantization of the cosmological constant must seem strange
    if you've never seen it before, but it's not really so weird.
    What's weird is that Witten is using Chern-Simons theory to
    determine the allowed values of the cosmological constant even
    though he wants to study what happens if gravity is *not*
    described by Chern-Simons theory!

    Witten knows this is weird: later he says "we used the gauge
    theory approach to get some hints about the right values of the
    cosmological constant simply because it was the only tool

    Indeed, the whole paper seems designed to refute the notion that
    mathematicians get less daring as they get older. He writes: "We
    make at each stage the most optimistic possible assumption."
    Perhaps he has some evidence for his guesses that he's not
    revealing yet. Or perhaps he's decided it takes courage verging
    on recklessness to track the Monster to its lair.

    Anyway: next Witten relates the level k to something called the
    "central charge" of the conformal field theory living at the
    boundary at infinity.

    What's the "central charge"? This is a standard concept in
    conformal field theory. Perhaps the simplest explanation is that
    in a conformal field theory, the total energy of the vacuum state
    is -c/24, where c is the central charge. So, naively you'd expect
    c = 0, but quantum effects make nonzero values of the vacuum energy
    possible, and even typical. A closely related cool fact is that
    the partition function of a conformal field theory is only a
    well-defined number up to multiples of

    exp(2 pi i c / 24)

    This means the partition function is a well-defined number when
    c is a multiple of 24. This happens in certain especially nice
    conformal field theories which are said to have "holomorphic

    The appearance of the magic number 24 here is the first taste of
    Monstrous Moonshine! For more on the importance of this number
    in string theory, see "week124", "week125" and "week126".

    As you can see, there are lots of subtleties here, which I really
    don't want to get into, but feel guilty about glossing over.

    Here's another. There are really *two* conformal field theories
    in this game: one that describes ripples of the gravitational
    field moving clockwise around the boundary at infinity, and
    another for ripples moving counterclockwise. Our simplifying
    assumption about left-right symmetry lets us describe these
    "right-movers" and "left-movers" with the same theory. So, both
    have the same central charge.

    In this case, the relation between central charge and level
    is simple:

    c = 24 k

    Next, Witten considers the situation where k takes its smallest
    interesting value: k = 1, so c = 24. It just so happens that
    c = 24 conformal field theories with holomorphic factorization
    have been classified, at least modulo a certain conjecture:

    2) A. N. Schellekens, Meromorphic c = 24 conformal field theories,
    Comm. Math. Phys. 153 (1993) 159-196. Also available as

    It's believed there are 71 of them. Which one could describe
    3d quantum gravity?

    Of these 71, ALL BUT ONE have gauge symmetries! Now, Witten
    is assuming 3d quantum gravity is *not* described by Chern-Simons
    theory, which is a gauge theory. So, he guesses that the one
    exceptional theory is the right one!

    And this is a very famous conformal field theory. It's a theory
    of a bosonic string wiggling around in a 26-dimensional spacetime
    curled up in clever way with the help of a 24-dimensional lattice
    called the Leech lattice. This theory is famous because its
    symmetry group is the Monster group! It is, in fact, the simplest
    thing we know that has the Monster group as symmetries.

    For more details, try these - in rough order of increasing

    3) Terry Gannon, Postcards from the edge, or Snapshots
    of the theory of generalised Moonshine, available as

    Terry Gannon, Monstrous Moonshine: the first twenty-five years,
    available as arXiv:math/0402345.

    4) Richard Borcherds, online papers, available at

    5) Igor Frenkel, James Lepowsky, Arne Meurman, Vertex operator
    algebras and the Monster, Academic Press, New York, 1988.

    Now, if this monstrous conformal field theory turned out to be
    3d quantum gravity in disguise - viewed from infinity, so to
    speak - it might someday give a much better understanding of
    Monstrous Moonshine. However, Witten gives no explanation as to
    *why* this theory should be 3d gravity, except for the indirect
    argument I just sketched. The precise relation between 3d
    quantum gravity and the bosonic string wiggling around in
    26 dimensions remains obscure.

    However, while Witten leaves this mysterious, he does offer a
    tantalizing extra tidbit of evidence that the relation is real!

    The partition function of the monstrous conformal field theory
    I just mentioned is the j-function, or more precisely:

    J(q) = q^{-1} + 196884 q + 21493760 q^2 + ...

    As I mentioned, this function shows up naturally in complex
    analysis. More precisely, it parametrizes the moduli space of
    elliptic curves (see "week125"). But, its bizarre coefficients
    turn out to be dimensions of interesting representations of the
    Monster group. For example, the smallest nontrivial representation
    of the Monster has dimension 196883; adding the trivial
    representation, we get 196884. This was one of several strange
    clues leading to the discovery of Monstrous Moonshine.

    What Witten does is assume that the monstrous conformal field
    theory describes 3d quantum gravity for k = 1, and then use
    properties of the j-function to compute the entropy of black

    I won't attempt to explain the calculation. Suffice it to say
    that the lightest possible black hole turns out to have
    196883 quantum states - its space of states is the smallest
    nontrivial representation of the Monster group. So, its entropy

    ln(196883) ~ 12.19

    On the other hand, Hawking's semiclassical calculation gives

    4 pi ~ 12.57

    The match is not perfect - but it doesn't need to be, since we
    expect quantum corrections to Hawking's formula for small black

    What's more impressive is that Witten can guess the entropy of
    the lightest possible black hole for other values of k - meaning,
    other values of the cosmological constant. The space of states
    of these black holes are always representations of the Monster
    group, so we get logarithms of weird-looking integers. For
    example, for k = 4 the entropy is

    ln(81026609426) ~ 25.12

    while Hawking's formula gives

    8 pi ~ 25.13

    Much better! And, using a formula of Petersson and Rademacher
    for asymptotics of coefficients of the j-function, together with
    some facts about Hecke operators, he shows that as k -> infinity,
    the agreement becomes perfect!

    In short, there are some fascinating hints of a relation between
    the Monster group and black hole entropies in 3d gravity, but the
    details of Witten's hoped-for "AdS-CFT correspondence" between 3d
    gravity and the monstrous conformal field theory remain obscure.
    Indeed, there are lots of problems with Witten's proposal:

    6) Jacques Distler, Witten on 2+1 gravity,

    But, time will tell. In fact, if history is any guide, we can
    expect to see armies of string theorists marching into this
    territory any day now. So, I'll just pose one question.

    There's a well-known route from 2d rational conformal field
    theories (or "RCFTs") to 3d topological quantum field theories
    (or "TQFTs"), which passes through modular tensor categories.
    For example, an RCFT called the Wess-Zumino-Witten model gives
    the TQFT called Chern-Simons theory.

    But, now Witten is saying 3d quantum isn't Chern-Simons theory;
    instead, it's something related to the monstrous CFT.

    So: is the monstrous conformal field theory known to be an RCFT?
    If so, what 3d TQFT does it give? Could this TQFT be the 3d
    quantum gravity theory Witten is seeking?

    Even though Witten is now claiming 3d quantum gravity *can't be*
    a TQFT, I think this is an interesting question. At the very
    least, I'd like to know more about this "Monster TQFT" - if it

    Now let's move from 3d quantum gravity to real-world particle

    Last week I described some mathematical relations between the
    Standard Model of particle physics, the most famous grand unified
    theories, and some "exceptional" structures in mathematics: the
    exceptional Lie group E6, the complexified octonionic projective
    plane, and the exceptional Jordan algebra.

    This week I want to go a bit further, and talk about the work
    of Kac, Larsson and others on the exceptional Lie superalgebras
    E(3|6), E(3|8) and E(5|10).

    As before, my goal is to point out some curious relations between
    the messy pack of particles we see in nature and the "exceptional"
    structures we find in mathematics. By this, I mean structures
    that show up when you classify algebraic gadgets, but don't fit
    into nice systematic infinite families. Right now the Monster
    is the king of all exceptional structures, the biggest of the
    26 sporadic finite simple groups. But, there are lots of other
    such structures, and they all seem to be related.

    As mentioned back in "week66", Edward Witten once suggested that
    the correct theory of our universe could be an exceptional
    structure of some sort. There's even a fun hand-wavy argument
    for this idea. It goes like this: the theory of our universe
    *must* be incredibly special, since out of all the theories we can
    write down, only one describes the universe that actually exists!

    In particular, lots of very simple theories do *not* describe our
    universe. So there must be some principle besides simplicity that
    picks out the theory of our universe.

    Unfortunately, when we try to think about these issues seriously,
    we're quickly led into very deep waters. In practice, people
    quickly muddy these waters and create a quagmire. It's very hard
    to discuss this stuff without uttering nonsense. If you want
    to see my try, look at "week146".

    But right now, I prefer to act like a sober, serious mathematical
    physicist. So, I'll tell you a bit about exceptional Lie
    superalgebras and how they could be related to the Standard Model.

    First, some history. In 1887, Wilhelm Killing sent a letter to
    Friedrich Engel saying he'd classified the simple Lie algebras.
    Besides the "classical" ones - namely the infinite series sl(n,C),
    so(n,C) and sp(n,C) - he found 6 exceptions: a 14-dimensional one,
    two 52-dimensional ones, a 78-dimensional one, a 133-dimensional
    one and a 248-dimensional one.

    In 1894, Eli Cartan finished a doctoral thesis in which he cleaned
    up Engel's work. In the process, he noticed that Engel's two
    52-dimensional Lie algebras were actually the same. Whoops!

    So, we now have just 5 "exceptional" simple Lie algebras. In
    order of increasing size, they're called G2, F4, E6, E7 and E8.

    In 1914, Cartan realized that the smallest exceptional Lie algebra,
    G2, comes from the symmetry group of the octonions! Later it was
    realized that all 5 are connected to the octonions. I've written
    a lot about this in previous Weeks, but most of that material can
    be found here:

    6) John Baez, Exceptional Lie algebras,

    Now, whenever mathematicians do something fun, they want to keep
    doing it, which means *generalizing* it.

    One way to generalize Cartan's work is to study "symmetric spaces",
    which I defined last week. Briefly, a symmetric space is a
    manifold equipped with a geometrical structure that's very
    symmetrical: so much so that every point is just like every other,
    and the view in any direction is the same as the view in the
    opposite direction.

    In fact, it was Cartan himself who invented the concept of
    symmetric space, and after he classified the simple Lie algebras
    he went ahead and classified these.

    More precisely, I think he classified the "compact Riemannian"
    symmetric spaces. Every simple Lie algebra gives one of these,
    namely a compact simple group. But, there are others too. So,
    compact Riemannian symmetric spaces are a nice generalization
    of simple Lie algebras - and I believe Cartan succeeded in
    classifying them all.

    Again, there are some infinite series, but also some exceptions
    coming from the octonions. I talked about one of these last week,
    namely EIII, the complexified octonionic projective plane. You
    can see a list here:

    7) Wikipedia, Riemannian symmetric space,

    For a quick intro to the classification of simple Lie algebras and
    compact Riemannian symmetric spaces, try this great book:

    8) Daniel Bump, Lie Groups, Springer, Berlin, 2004.

    For a slower, more thorough introduction, try the book by Helgason
    mentioned last Week.

    A second way to generalize Cartan's work is to consider simple
    Lie *superalgebras*.

    Lie superalgebras are just like Lie algebras, except they're split
    into an "even" or bosonic and "odd" or fermionic part. The idea
    is that we stick minus signs in the usual Lie algebra formulas
    whenever we switch two "odd" elements.

    This is very natural from a physics viewpoint, since whenever
    you switch two identical fermions, the wavefunction of the
    universe gets multiplied by -1. (Take my word for it - I've seen
    it happen!)

    It's also very natural from a math viewpoint, since "super vector
    spaces" form a symmetric monoidal category with almost all the
    nice properties of plain old vector spaces. This lets crazed
    mathematicians and physicists systematically generalize pretty
    much all of linear algebra to the "super" world. So, why not Lie

    The simple Lie superalgebras were classified by Victor Kac in

    9) Victor Kac, Lie superalgebras, Adv. Math. 26 (1977), 8-96.

    Not counting the ordinary simple Lie algebras, there are 8
    series of simple Lie superalgebras and a few exceptional ones.
    At least some of these exceptions come from the octonions:

    10) Anthony Sudbery, Octonionic description of exceptional
    Lie superalgebras, Jour. Math. Phys. 24 (1983), 1986-1988.

    Do they all? I don't know! Someone please tell me!

    A third way to generalize Cartan's work is to classify
    *infinite-dimensional* simple Lie algebras - or for that matter,
    Lie superalgebras.

    So far I've implicitly assumed all our algebraic gadgets are
    finite-dimensional, but we can lift that restriction. If you
    try to classify infinite-dimensional gadgets without *any*
    restrictions, it can get really hairy. It turns out the nice
    thing is to classify "linearly compact" infinite-dimensional
    simple Lie algebras. I won't define the quoted phrase, since
    it's technical and it's explained near the beginning
    of this paper:

    11) Victor Kac, Classification of infinite-dimensional simple
    linearly compact Lie superalgebras, Erwin Schroedinger Institut
    preprint, 1998.
    Available at http://www.esi.ac.at/Preprint-shadows/esi605.html

    Anyway, back in 1880 Lie himself made a guess about
    infinite-dimensional Lie algebras that would solve the problem
    I'm talking about now, though he didn't phrase it in the modern
    way. And, Cartan proved Lie's guess in 1909! Actually, there
    was a hole in Cartan's proof, which was only noticed much later.
    It was filled by Guillemin, Quillen and Sternberg in 1966.

    So, here's the answer: there are 4 families of linearly compact
    infinite-dimensional simple Lie algebras, and no exceptions.
    Ignoring an important nuance I'll explain later, these are:

    A) The Lie algebra of all complex vector fields on C^n.

    B) The Lie algebra of all complex vector fields v on C^n
    that are "divergence-free":

    div v = 0

    C) The Lie algebra of all complex vector fields v on C^{2n}
    that are "symplectic":

    L_v omega = 0

    where omega is the usual symplectic structure on C^{2n},
    and L means "Lie derivative"

    D) The Lie algebra of all complex vector fields v on C^{2n+1}
    that are "contact":

    L_v alpha = f alpha

    for some function f depending on v, where alpha is the usual
    contact structure on C^{2n+1}.

    If you don't know about symplectic structures or contact
    structures, don't worry - we won't need them now. The main point
    is that they're differential forms that show up throughout
    classical mechanics. So, this classification theorem is
    surprisingly nice.

    Notice: no exceptions! That's a kind of exception in its own

    In 1998, Victor Kac proved the "super" version of this result.
    In other words, he classified linearly compact infinite-dimensional
    Lie superalgebras! This result is Theorem 6.3 of his paper above.
    There turn out to be 10 families and 6 exceptions, which are
    called E(1|6), E(2|2), E(3|6), E(3|8), E(4|4) and E(5|10).

    Many of the families are straightforward "super" generalizations
    of the 4 families I just showed you. Some are stranger. Most
    important for us today are the exceptions discovered by Irina
    Shchepochkina in 1983:

    12) Irena Shchepochkina, New exceptional simple Lie superalgebras,
    C. R. Bul. Sci. 36 (1983), 313-314.

    The easiest to explain is E(5|10). And, you'll soon see that the
    number 5 here is related to the math of the SU(5) grand unified
    theory, which I explained last Week!

    The even part of E(5|10) is the Lie algebra of divergence-free complex
    vector fields on C^5.

    The odd part of E(5|10) consists of closed complex 2-forms on C^5.

    The bracket of two even guys is the usual Lie bracket of vector

    The bracket of an even guy and an odd guy is the usual "Lie
    derivative" of a differential form with respect to a vector field.

    The only tricky bit is the bracket of two odd guys! So, suppose
    mu and nu are closed complex 2-forms on C^5. Their wedge product
    is a 4-form mu ^ nu. But, we can identify this with a vector
    field v by demanding:

    i_v vol = mu ^ nu

    Here vol is the volume form:

    vol = dx_1 ^ dx_2 ^ dx_3 ^ dx_4 ^ dx_5

    and i_v vol is the "interior product", which feeds v into vol and
    leaves us with a 4-form. You can check that this vector field v
    is divergence-free. So, we define the bracket of mu and nu to be

    sl(5,C) sits inside the even part of E(5|10) in a nice way, as
    the divergence-free vector fields whose coefficients are *linear*
    functions on C^5. So, since su(5) sits inside sl(5,C), we get a
    tempting relation to SU(5).

    (Now I'll come clean now and explain the "important nuance" I
    ignored earlier. For the classification theorems I mentioned
    earlier, we must use vector fields and differential forms with
    *formal power series* as coefficients. But for the purposes
    of mathematical physics, we should keep a more flexible attitude.)

    Next, what about E(3|6)? This is contained in E(5|10). To define
    it, we give E(5|10) a clever grading where x_1, x_2, x_3 are
    treated differently from the other two variables. Then we take
    the subalgebra of degree-zero guys. The details are explained in
    the above papers - or more simply, here:

    13) Victor Kac, Classification of infinite-dimensional simple
    groups of supersymmetries and quantum field theory, available as

    All this is reminiscent of how SU(5) contains the gauge group of
    the Standard Model, namely S(U(3) x U(2)). In particular, the
    even part of E(3|6) contains the Lie algebra

    sl(3,C) + sl(2,C) + gl(1,C)

    in a canonical way. So, any representation of E(3|6) automatically
    gives a representation of the Standard Model Lie algebra

    su(3) + su(2) + u(1)

    And in the above paper Kac goes even further! He defines a fairly
    natural class of representations of E(3|6), and proves something
    remarkable: these restrict to representations of

    su(3) + su(2) + u(1)

    that correspond precisely to the gluon, the photon and the W and
    Z bosons, and the quarks and leptons in one generation...

    ... together with one other particle, which is *not* the Higgs
    boson, but instead acts like a gluon with electric charge +-1.


    One nice thing is how these Lie superalgebras get both bosons and
    fermions into the game in a natural way without forcing the
    existence of a bunch of unseen "superpartners". One unfortunate
    thing is that the above result gives no hint as to why there
    should be three generations of quarks and leptons. However, Kac
    and Rudakov develop some mathematics to address that question

    14) Victor Kac and Alexi Rudakov, Representations of the
    exceptional Lie superalgebra E(3,6): I. Degeneracy conditions.
    Available as math-ph/0012049.

    Representations of the exceptional Lie superalgebra E(3,6): II.
    Four series of degenerate modules. Available as math-ph/0012050.

    Representations of the exceptional Lie superalgebra E(3,6)
    III: Classification of singular vectors. Available as

    Their results are summarize at the end of this review article:

    15) Victor Kac, Classification of supersymmetries, Proceedings
    of the ICM, Beijing, 2002, vol. 1, 319-344. Available as

    Here Kac writes that "three generations of leptons occur in
    the stable region [whatever that means], but the situation with
    quarks is more complicated: this model predicts a complete fourth
    generation of quarks and an incomplete fifth generation (with
    missing down type triplets)."

    So, while I don't understand "this model", it seems tantalizingly
    close to capturing the algebraic patterns in the Standard Model...
    without quite doing so.

    Some more nice explanations and references can be found here:

    16) Irina Shchepochkina, The five exceptional simple Lie
    superalgebras of vector fields. Available as hep-th/9702.3121.

    17) Pavel Grozman, Dimitry Leites and Irina Shchepochkina,
    Defining relations for the exceptional Lie superalgebras of
    vector fields pertaining to The Standard Model, available as

    18) Pavel Grozman, Dimitry Leites and Irina Shchepochkina,
    Invariant operators on supermanifolds and Standard Models,
    available as math.RT/0202193.

    Thomas Larsson has been working on similar ideas, mainly
    using E(3|8) instead. This also contains

    sl(3,C) + sl(2,C) + gl(1,C)

    in a canonical way.

    19) Thomas A. Larsson, Symmetries of everything, available as

    Exceptional Lie superalgebras, invariant morphisms, and a
    second-gauged Standard Model, available as math-ph/020202.

    Thomas A. Larsson, Maximal depth implies su(3)+su(2)+u(1),
    available as hep-th/0208185.

    Alas, E(3|8) gets the hypercharges of some fermions wrong.
    Larsson seems to say this problem also occurs for E(3|6),
    which would appear to contradict what Kac claims - but I
    could be misunderstanding.

    I'll end with few questions. First, is there any relation
    between the exceptional Lie superalgebras E(5|10), E(3|6) or
    E(5|10) and the exceptional Lie algebra e6? Last week I
    explained some relations between e6 and the Standard Model;
    are those secretly connected to what I'm discussing this week?

    Second, has anyone tried to unify all three generalizations of
    Cartan's classification of simple Lie algebras? Starting from
    simple Lie algebras, we've seen three ways to generalize:

    A) go to symmetric spaces,
    B) go the "super" version,
    C) go to the infinite-dimensional case.

    So: has anyone tried to classify infinite-dimensional super
    versions of symmetric spaces? Or even finite-dimensional ones?

    (Maybe the super version of a symmetric space should be
    called a "supersymmetric space", just for the sake of a nice pun.)

    Next, the Tale of Groupoidification! I'll keep this week's
    episode short, since you're probably exhausted already.

    I want to work my way to the concept of "Hecke operator"
    through a series of examples. The examples I'll use are a bit
    trickier than the concept I'm really interested in, since the
    examples involve integrals, where the Hecke operators I ultimately
    want to discuss involve sums. But, the examples are nice if
    you like to visualize stuff...

    In these examples we'll always have a relation between two sets
    X and Y. We'll use this to get an operator that turns functions
    on X into functions on Y - a "Hecke operator".

    A) The Radon transform in 2 dimensions

    Suppose you're trying to do a CAT scan. You want to obtain
    a 3d image of someone's innards. Unfortunately, all you do is
    take lots of 2d X-ray photos of them. How can you assemble all
    this information into the picture you want?

    Who better to help you out than a guy named after a
    radioactive gas: Radon!

    In 1917, the Viennese mathematician Johann Radon tackled a related
    problem one dimension down. You could call it a "CAT scan for

    Suppose you want to obtain a complete image of the insides of a
    2-dimensional person, but all you can do is shine beams of X-rays
    through them and see how much each beam is attenuated.

    So, mathematically: you have a real-valued function on the plane -
    roughly speaking, the density of your flatlander. You're trying
    to recover this function from its integrals along all possible
    lines. Someone hands you this function on the space of *lines*,
    and you're trying to figure out the original function on the space
    of *points*.

    (Points lying on lines! If you've been following the Tale of
    Groupoidification, you'll know this "incidence relation" is
    connected to Klein's approach to geometry, and ultimately to
    spans of groupoids. But pretend you don't notice, yet.)

    Now, it's premature to worry about this tricky "inverse problem"
    before we ponder what it's the inverse of: the "Radon transform".
    This takes our original function on the space of *points* and
    gives a function on the space of *lines*.

    Let's call the Radon transform T. It takes a function f on the
    space of points and gives a function Tf on the space of lines,
    as follows. Given a line y, (Tf)(y) is the integral of f(x) over
    the set of all points x lying on y.

    What Radon did is figure out a nice formula for the inverse of
    this transform. But that's not what I'm mainly interested in now.
    It's the Radon transform itself that's a kind of Hecke operator!

    Next, look at another example.

    B) The X-ray transform in n dimensions

    This is an obvious generalization to higher dimensions of what
    I just described. Before we had a space

    X = {points in the plane}

    and a space

    Y = {lines in the plane}

    and an incidence relation

    S = {(x,y): x is a point lying on the line y}

    If we go to n dimensions, we can replace all this with

    X = {points in R^n}

    Y = {lines in R^n}

    S = {(x,y): x is a point lying on the line y}

    Again, the X-ray transform takes a function f on the space of
    points and gives a function Tf on the space of lines. Given a
    line y, (Tf)(y) is the integral of f(x) over the set of all x
    with (x,y) in S.

    Next, yet another example!

    C) The Radon transform in n dimensions

    Radon actually considered a different generalization of the
    2d Radon transform, using hyperplanes instead of lines. Using
    hyperplanes is nicer, because it gives a very simple relationship
    between the Radon transform and the Fourier transform. But never
    mind - that's not the point here! The point is how similar
    everything is. Now we take:

    X = {points in R^n}

    Y = {hyperplanes in R^n}

    S = {(x,y): x is a point lying on the hyperplane y}

    And again, the Radon transform takes a function f on X
    and gives a function Tf on Y. Given y in Y, (Tf)(y) is the
    integral of f(x) over the set of all x with (x,y) in S.

    We're always doing the same thing here. Now I'll describe
    the general pattern a bit more abstractly.

    We've always got three spaces, and maps that look like this:

    / \
    / \
    P/ \Q
    / \
    v v
    X Y

    In our examples so far these maps are given by

    P(x,y) = x
    Q(x,y) = y

    But, they don't need to be.

    Now, how do we get a linear operator in this situation?

    Easy! We start with a real-valued function on our space X:

    f: X -> R

    Then we take f and "pull it back along P" to get a function on S.
    "Pulling back along P" is just impressive jargon for composing
    with P:

    f o P: S -> R

    Next, we take f o P and "push it forwards along Q" to get a
    function on Y. The result is our final answer, some function

    Tf: Y -> R

    "Pushing forwards along Q" is just impressive jargon for
    integrating: Tf(y) is the integral over all s in S with Q(s) = y.
    For this we need a suitable measure, and we need the integral
    to converge.

    This is the basic idea: we define an operator T by pulling
    back and then pushing forward functions along a "span", meaning
    a diagram shaped like a bridge:

    / \
    / \
    P/ \Q
    / \
    v v
    X Y

    But, the reason this operator counts as a "Hecke operator"
    is that it gets along with a symmetry group G that's acting
    on everything in sight. In the examples so far, this is
    the group of Euclidean symmetries of R^n. But, it could be

    This group G acts on all 3 spaces: X, Y, and S. This makes the
    space of functions on X into a representation of G! And, ditto
    for the space of function on Y and S.

    Furthermore, the maps P and Q are "equivariant", meaning

    P(gs) = gP(s)


    Q(gs) = gQ(s)

    This makes "pulling back along P" into an intertwining operator
    between representations of G. "Pushing forwards along Q" will
    also be an intertwining operator if the measure we use is
    G-invariant in a suitable sense. In this case, our transform T
    becomes an intertwining operator between group representations!
    Let's call an intertwining operator constructed this way a "Hecke

    If you're a nitpicky person, e.g. a mathematician, you may wonder
    what I mean by "a suitable sense". Well, each "fiber" Q^{-1}(y)
    of the map

    Q: S -> Y

    needs a measure on it, so we can take a function on S and
    integrate it over these fibers to get a function on Y. We need
    this family of measures to be invariant under the action of G,
    for pushing forwards along Q be an intertwining operator.

    This stuff about invariant families of measures is mildly
    annoying, and so is the analysis involved in making precise
    *which* class of functions on X we can pull back to S and then
    push forward to Y - we need to make sure the integrals converge,
    and so on. When I really get rolling on this Hecke operator
    business, I'll often focus on cases where X, Y, and S are
    *finite* sets... and then these issues go away!

    Hmm. I'm getting tired, but I can't quit until I say one more
    thing. If you try to read about Hecke operators, you *won't*
    see anything about the examples I just mentioned! You're most
    likely to see examples where X and Y are spaces of lattices in
    the complex plane. This is the classic example, which we're
    trying to generalize. But, this example is more sophisticated
    than the ones I've mentioned, in that the "functions" on X and
    Y become "sections of vector bundles" over X and Y. The same
    sort of twist happens when we go from the Radon transform to the
    more general "Penrose transform".

    Anyway, next time I'll talk about some really easy examples,
    where X, Y, and S are finite sets and G is a finite group.
    These give certain algebras of Hecke operators, called "Hecke

    In the meantime, see if you can find *any* reference in the
    literature which admits that "Hecke algebras" are related
    to "Hecke operators". It ain't easy!

    It's a great example of a mathematical cover-up - one we're
    gonna bust wide open.


    Addendum: for more discussion, go to the n-Category Cafe:


    Previous issues of "This Week's Finds" and other expository articles on
    mathematics and physics, as well as some of my research papers, can be
    obtained at


    For a table of contents of all the issues of This Week's Finds, try


    A simple jumping-off point to the old issues is available at


    If you just want the latest issue, go to

  2. jcsd
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Can you offer guidance or do you also need help?
Draft saved Draft deleted