Newsgroup sci.math.research 6111

Directory

Subject: Re: Looking for E s.t. Pi1(E)=S1 -- From: greg@math.math.ucdavis.edu (Greg Kuperberg)
Subject: Re: a question about groups generated by a single conjugacy class -- From: greg@math.math.ucdavis.edu (Greg Kuperberg)
Subject: Re: terminology -- From: Ray DeCampo
Subject: diffeomorphism -- From: pcascini@hotmail.com (paolo cascini)
Subject: Re: Looking for E s.t. Pi1(E)=S1 -- From: baez@math.ucr.edu (john baez)
Subject: Re: Interesting DE Problem!!! -- From: achim@mipool.uni-jena.de (Achim Flammenkamp)
Subject: Re: Looking for E s.t. Pi1(E)=S1 -- From: ctm@maize.berkeley.edu (C. T. McMullen)
Subject: This Week's Finds in Mathematical Physics (Week 93) -- From: baez@math.ucr.edu (John Baez)
Subject: Poincare series for the depths of roots in a root system -- From: dima@win.tue.nl (Dmitrii V. Pasechnik)
Subject: Galois Field Problem -- From: Amr Mohamed Youssef
Subject: Re: surface and curve intersection algorithm -- From: Saul Youssef
Subject: Re: diffeomorphism -- From: Pat.McSwiggen@UC.Edu (Patrick D. McSwiggen)
Subject: Re: This Week's Finds in Mathematical Physics (Week 93) -- From: baez@math.ucr.edu (john baez)
Subject: Re: Interesting DE Problem!!! -- From: baez@math.ucr.edu (john baez)
Subject: Re: terminology -- From: ikastan@alumnae.caltech.edu (Ilias Kastanas)
Subject: infinitely near points -- From: adler@pulsar.cs.WKU.EDU (Allen Adler)

Articles

Subject: Re: Looking for E s.t. Pi1(E)=S1
From: greg@math.math.ucdavis.edu (Greg Kuperberg)
Date: 25 Oct 1996 23:57:05 GMT
In article <54qsa1$bcf@nef.ens.fr> mercat@clipper.ens.fr (Christian Mercat) writes:
>I would like to understand why the circle S1 should be the
>complexified of Z2={-1,+1}. For that, I should find "something", a CW
>complex, a manifold, (infinite dimensional), E such that S1=Pi1(E).
Some mathematicians have a philosophy to the effect that many real
structures have complex analogues which differ non-trivially.
Obviously the group +-1 is the set of unit elements in R and
correspondingly S^1 is the set of unit elements in C.  Extending this
parallel, RP^n is a partial classifying space of +-1 and CP^n is a
partial classifying space of the circle.
To further extend the parallel, Stiefel-Whitney classes (which are
Z/2-valued) give you obstructions to trivializing a real bundle over a
topological space, and Chern classes (which are Z-valued, where Z =
pi_1 of the same old circle) give you obstructions to trivializing a
complex bundle.  Indeed the top Stiefel-Whitney class of the tangent
bundle of a manifold tells you whether the manifold is orientable, and
you can interpret the top Chern class as the obstruction to "complex
orientability".  So you can conclude that among Kahler manifolds, those
that are Calabi-Yau are "complex orientable".
If you give CP^n the usual topology that makes it a manifold, then it
is a partial classifying space of the circle as a Lie group rather than
as a discrete group.  But there is a second topology on CP^n such that
pi_1(CP^2) is already a circle.  As you know, CP^2 is the set of
triples of complex numbers (z_1,z_2,z_3), not all zero, modulo the
equivalence given by multiplying a complex number.  Now give the
original set of triples of complex numbers a topology such that
(z_1(t),z_2(t),z_3(t)) is continuous if and only if it is continuous in
the usual sense and each z_i(t) does not vary in phase except when it
passes through 0.  This is your desired space E.
-- 
   /\   Greg Kuperberg        greg@math.ucdavis.edu
  /  \ 
  \  /  Recruiting or seeking a job in math?  Check out my Generic Electronic
   \/   Job Application form, http://www.math.ucdavis.edu/~greg/geja/
Return to Top
Subject: Re: a question about groups generated by a single conjugacy class
From: greg@math.math.ucdavis.edu (Greg Kuperberg)
Date: 26 Oct 1996 00:25:50 GMT
Two weeks ago, Lee Rudolph asked about the word length of elements of
the braid group if the generators an entire conjugacy class, in
particular the conjugacy class of half twists of a pair of strands.  I
misread the question at the time, but to atone for my sins I went and
asked Bill Thurston.  It has led to some interesting joint research.
In general new research seems to be a remarkably easy prospect when
talking to Thurston.
Here is an argument that he found that addresses the original
question.  Theorem:  Many elements g in the commutator subgroup have
the property that the commutator length of g^n grows linearly in n.
This implies that the conjugacy word length also grows linearly.
The proof goes like this.  Instead of the braid group on n strands,
consider the quotient mapping class group of a sphere with n+1 marked
points.  This mapping class group has a Teichmuller space which is
negatively curved in the Weil-Peterson metric.  Take some closed
geodesic gamma in the moduli space, a quotient orbifold, which does not
simply double back along a geodesic line segment, and let x be a
corresponding group element.  (I.e., take a hyperbolic element x of the
mapping class group which is not contained in a Z/2 semidirect Z.) Then
if x has commutator length g, gamma bounds a mapped-in surface S of
genus g, and vice-versa.  The surface that is mapped in is not allowed
to cheat at the orbifold points.  Now triangulate S economically and
straighten each triangle in the target in the fashion of Gromov.  Since
Weil-Peterson is geodesically convex, it is possible to do this.
Although Weil-Peterson has regions where negative curvature approaches
0, the convex hull of gamma avoids these, so that after straightening,
you get an upper bound on the area of S which is linear in g using a
standard relation between negative curvature and area of straight
triangles.  On the other hand, if you pick a 1-form alpha whose line
integral on gamma is non-zero, you get a lower bound on the area of S
by integrating d alpha and using Stokes' theorem.  Replacing gamma by
gamma^n, this lower bound grows linearly.
-- 
   /\   Greg Kuperberg        greg@math.ucdavis.edu
  /  \ 
  \  /  Recruiting or seeking a job in math?  Check out my Generic Electronic
   \/   Job Application form, http://www.math.ucdavis.edu/~greg/geja/
Return to Top
Subject: Re: terminology
From: Ray DeCampo
Date: Fri, 25 Oct 1996 23:05:19 -0400
Michael Hardy wrote:
> 
>         Last week I wrote:
> 
> >         When, as in defining quotient groups, one defines an operation on
> > classes by picking a member of each class to be operated on, and performing
> > an operation on the members, (and then worrying about "problematic", (i.e.
> > possibly ambiguous and possibly not) definitions), then one is defining an
> > operation *******-wise.  I'm looking for this locution, filling in the
> > row of asterisks with a one- or two-syllable word.  Does anyone know it?
> 
>         In a post that hasn't yet appeared at this site
>         (I got it by gopher-ing to math.lfc.edu
>         Andrew Stacey  replied:
> 
> > This is a complete stab in the dark, it's probably not recognised
> > terminology but if you just want something that carries the idea try:
> >
> > representative-wise
> >
> > Since the 'x' in x+H is a representative element of the equivalence class
> > defined by: x~x' iff x-x' is in H (using addition rather than mult as
> > subtraction is easier to type than inverses!)
> [cut critique of 'representative-wise']
How about dropping the whole '-wise' suffix?  How about saying that the 
operation is defined via representatives?  (or 'on representatives' or 
{your favorite preposition here} representatives?)
Ray DeCampo
Return to Top
Subject: diffeomorphism
From: pcascini@hotmail.com (paolo cascini)
Date: 26 Oct 1996 10:30:14 +0100
I'm not English so you have to soory my language...
can you find two space that are diffeomorph of class Cn for each n, but 
that are not diffepmorph of class c-inf.
Return to Top
Subject: Re: Looking for E s.t. Pi1(E)=S1
From: baez@math.ucr.edu (john baez)
Date: 26 Oct 1996 14:21:08 -0700
In article <54qsa1$bcf@nef.ens.fr>,
Christian Mercat  wrote:
>I would like to understand why the circle S1 should be the complexified
>of Z2={-1,+1}. For that, I should find "something", a CW complex, a manifold,
>(infinite dimensional), E such that S1=Pi1(E).
>And after, if that E is by any chance the complexified of something Er, I would
>love that Z2=Pi0(Er).
For any group G there is an obvious "best" space whose fundamental
group is G, namely the Eilenberg-MacLane space KG.  This has pi_1(KG) = G
and pi_n(KG) = 0 for all higher n.  It's unique up to homotopy equivalence,
which is as unique as things should be in homotopy theory.
For G = Z_2 you get KG = RP^infinity, the infinite-dimensional real projective
space.   
How about for S^1?
As far as I know, you're not going to get the all-important *topological
group* structure of S^1 simply by taking the fundamental group of something.  
So it makes sense to switch directions a little bit here and look at something 
called "BG" instead of KG.  When G is a discrete topological group, BG is
just the same as KG defined as above.  But more generally...
For any decent topological group G, BG is the classifying space for G bundles: 
i.e., there is a "universal" G-bundle over BG, such any G-bundle over any decent
space X is isomorphic to the pullback of this bundle by some map f: X -> BG, 
and homotopy classes of maps f: X -> BG are in 1-1 correspondence with isomorphism 
classes of G-bundles over X.  
Giving Z_2 the structure of a discrete group, for G = Z_2 we get 
BG = RP^infinity as before.  And --- lo and behold --- when G = S^1 we
get BG = CP^infinity, the infinite-dimensional complex projective space.  
I think this is a good way to make precise what you want to be true.  
If one stares at the proof, it really does just follow from the fact 
that S1 is U(1), while Z_2 is O(1).  In fact, the same thing happens in 
fancier situations: the classifying space of the unitary group U(n) 
is the complexification of the classifying space of the orthogonal group 
O(n).  
Similar things hold for quaternions and the group Sp(n), so you might
want to ponder the notion of "quaternionification" while you're at it.
The above stuff may sound fancy but, to the extent that it's correct,
it can be found in any good book on algebraic topology, like Spanier's
"Algebraic Topology".  You might also peek at Adam's "Infinite Loop
Spaces" for more on this KG stuff.  I think I need to reread it myself,
now!
Return to Top
Subject: Re: Interesting DE Problem!!!
From: achim@mipool.uni-jena.de (Achim Flammenkamp)
Date: 27 Oct 1996 15:41:43 GMT
In article <54f3e4$gh6@amanda.dorsai.org>,
Abe Mantell  wrote:
>
>Hello,
>
>I have been quite a bit baffled by the following initial value problem:
>
>y' = y^2 + x^2, y(0)=1
>
>The slope field gives no hint of any singularities, only that the solution
>"grows" rapidly for increasing x & y.  However, the analytic solution 
There are different kinds of singularities (in one independent variable). 
1) risable singularities
2) improper(sorry don't know the correct english terminology) singularites
3) proper singularities
1) These are trivial: like y=(x^2-1)/(x-1) has a singularity at x=1
2) these are simple: F(x) has an improper sing <=> \exist P(x): P(x)*F(x)
   has a Tylor-sum and P(x) is a polynom not identical zero
3) All others
The big different is 
1) are some how artifical
2) have a Laurant-sum (in an open set) with only finite many negative indices
3) have no Laurant-sum or this has infinit nonzero coefficients for the 
   terms with negative indices
Only the last are the "proper" ones which should/are recognizable at a DE.
These of case 2) can be covered by the DE.
In function theory, there is a theorem, that says that "the continous FORTSEZUNG
in singularites of type 1) and 2) ALWAYS exists unique (no wonder: multiply by P(x) :)
. This explains that the description of such a function in form of a DE 
could give you no evidence of those "polstellen" .
Excuse my bad english, but I had no dictonary by hand :)
Look at C if you want understand the behaviour of R-functions.
achim
--
Achim Flammenkamp   Fakultät für Mathematik und Informatik   Universität Jena
Ernst-Abbe-Platz 2-4, Bau 13  07743 Jena   Germany   Phone: +49 (03641) 6-38661
email: achim@mipool.uni-jena.de   URL: http://www.mathematik.uni-jena.de/~achim/
localtime: UT+2   768bit-PGP-Key ID DAB9090D   
-- 
Achim Flammenkamp   Fakult"at f"ur Mathematik und Informatik
Universit"at Jena    Leutragraben 1    07743 Jena    Federal Republic of
Germany    Phone: +49 (03641) 6-30817   e-mail: achim@mipool.uni-jena.de
localtime: GMT+1        < Do you play Go ?   I'm  Game of Life  and  Go  Fan >
Return to Top
Subject: Re: Looking for E s.t. Pi1(E)=S1
From: ctm@maize.berkeley.edu (C. T. McMullen)
Date: 27 Oct 1996 14:16:28 GMT
In article <54qsa1$bcf@nef.ens.fr>,
Christian Mercat  wrote:
>I would like to understand why the circle S1 should be the complexified
>of Z2={-1,+1}. For that, I should find "something", a CW complex, a manifold,
>(infinite dimensional), E such that S1=Pi1(E).
>And after, if that E is by any chance the complexified of something Er, I would
>love that Z2=Pi0(Er).
>
Per Manin's principle from your previous post, the complexification
of Z/2 = S_2 (the symmetric group on 2 symbols) should be
the braid group B_2 = Z, not the circle.
And this works nicely:  Z/2 = \pi_0(R^*) and Z = \pi_1(C^*),
where R and C are the real and complex numbers, and * denotes
the group of units.
Return to Top
Subject: This Week's Finds in Mathematical Physics (Week 93)
From: baez@math.ucr.edu (John Baez)
Date: 27 Oct 1996 18:21:18 -0800
This Week's Finds in Mathematical Physics - Week 93
John Baez
Lately I've been trying to learn more about string theory.  I've always
had grave doubts about string theory, but it seems worth knowing about.
As usual, when I'm trying to learn something I find it helpful to write
about it --- it helps me remember stuff, and it points out gaps in my
understanding.  So I'll start trying to explain some string theory in
this and forthcoming Week's Finds.
However: watch out!  This isn't going to be a systematic introduction to
the subject.  First of all, I don't know enough to do that.  Secondly,
it will be very quirky and idiosyncratic, because the aspects of string
theory I'm interested in now aren't necessarily the ones most string
theorists would consider central.  I've been taking as my theme of
departure, "What's so great about 10 and 26 dimensions?"  When one
reads about string theory, one often hears that it only works in 10
or 26 dimensions --- and the obvious question is, why?
This question leads one down strange roads, and one runs into lots of
surprising coincidences, and spooky things that sound like coindences
but might NOT be coincidences if we understood them better.   
For example, when we have a string in 26 dimensions we can think of it
as wiggling around in the 24 directions perpendicular to the
2-dimensional surface the string traces out in spacetime (the "string
worldsheet").  So the number 24 plays an especially important role in
26-dimensional string theory.  It turns out that
                 1^2 + 2^2 + 3^2 + ... + 24^2 = 70^2.  
In fact, 24 is the only* integer n > 1 such that the sum of squares 
from 1^2 to n^2 is itself a perfect square.  Is this a coincidence?
Probably not, as I'll eventually explain!  This is just one of
many eerie facts one meets when trying to understand this stuff.
For starters I just want to explain why dimensions of the form 8k + 2 
are special.  Notice that if we take k = 0 here we get 2, the
dimension of the string worldsheet.  For k = 1 we get 10, the dimension
of spacetime in "supersymmetric string theory".  For k = 3 we get 26,
the dimension of spacetime in "purely bosonic string theory".  So these
dimensions are important.  What about n = k and the dimension 18, I hear
you ask?  Well, I don't know what happens there yet... maybe someone can
tell me!  All I want to do now is to explain what's good about
8n + 2.
But I need to start by saying a bit about fermions.  
Remember that in the Standard Model of particle physics --- the model
that all fancier theories are trying to outdo --- elementary particles
come in 3 basic kinds.  There are the basic fermions.  In general a
"fermion" is a particle whose angular momentum comes in units of
Planck's constant hbar times 1/2, 3/2, 5/2, and so on.  Fermions satisfy
the Pauli exclusion principle --- you can't put two identical fermions
in the same state.  That's why we have chemistry: the electrons stack up
in "shells" at different energy levels, instead of all going to the
lowest-energy state, because they are fermions and satisfy the exclusion
principle.  In the Standard Model the fermions go like this:
        LEPTONS                                     QUARKS
electron        electron neutrino         up quark          down quark
muon            muon neutrino             strange quark     charm quark
tauon           tauon neutrino            top quark         bottom quark
There are three "generations" here, all rather similar to each other.  
There are also particles in the Standard Model called "bosons" having
angular momentum in units of hbar times 0,1,2, and so on.  Identical
bosons, far from satisfying the exclusion principle, sort of like to all
get into the same state: one sees this in phenomena such as lasers,
where lots of photons occupy the same few states.  Most of the bosons
in the Standard Model are called "gauge bosons".  These carry the 
different forces in the standard model, by which the particles interact:
    ELECTROMAGNETIC FORCE          WEAK FORCE         STRONG FORCE
    photon                          W+                  6 gluons 
                                    W-
                                    Z  
Finally, there is also a bizarre particle in the Standard Model called the 
"Higgs boson".  This was first introduced as a rather ad hoc hypothesis:
it's supposed to interact with the forces in such a way as to break the
symmetry that would otherwise be present between the electromagnetic
force and the weak force.  It has not yet been observed; finding it would
would represent a great triumph for the Standard Model, while *not*
finding it might point the way to better theories.  
Indeed, while the Standard Model has passed many stringent experimental
tests, and successfully predicted the existence of many particles which
were later observed (like the W, the Z, and the charm and top quarks),
it is a most puzzling sort of hodgepodge.  Could nature really be this
baroque at its most fundamental level?  Few people seem to think so;
most hope for some deeper, simpler theory.
It's easy to want a "deeper, simpler theory", but how to get it?  What
are the clues?  What can we do?  Experimentalists certainly have their
work cut out for them.  They can try to find or rule out the Higgs.
They can also try to see if neutrinos, assumed to be massless in the
Standard Model, actually have a small mass --- for while the Standard
Model could easily be patched if this were the case, it would shed
interesting light on one of the biggest mysteries in physics, namely why
the fermions in nature seem not to be symmetric under reflection, or
"parity".  Right now, we believe that neutrinos only exist in a
left-handed form, rotating one way but not the other around the axis
they move along.  This is intimately related to their apparent
masslessness.  In fact, for reasons that would take a while to explain,
the lack of parity symmetry in the Standard Model forces us to assume
all fermions acquire their mass only through interaction with the Higgs
particle!  For more on the neutrino mass puzzle, try:
1) Paul Langacker, Implications of neutrino mass,
http://dept.physics.upenn.edu/~www/neutrino/jhu/jhu.html 
And, of course, experimentalists can continue to do what they always do
best: discover the utterly unexpected.  
Theorists, on the other hand, have been spending the last couple of
decades poring over the standard model and trying to understand what
it's telling us.  It's so full of suggestive patterns and partial 
symmetries!  First, why are there 3 forces here?  Each force goes along
with a group of symmetries called a "gauge group", and electromagnetism
corresponds to U(1), while the weak force corresponds to SU(2) and the
strong force corresponds to SU(3).  (Here U(n) is the group of n x n
unitary complex matrices, while SU(n) is the subgroup consisting of
those with determinant equal to 1.)  Well, actually the Standard Model
partially unifies the electromagnetic and weak force into the
"electroweak force", and then resorts to the Higgs to explain why these
forces are so different in practice.  Various "grand unified theories"
or "GUTs" try to unify the forces further by sticking the group SU(3) x
SU(2) x U(1) into a bigger group --- but then resort to still more
Higgses to break the symmetry between them!
Then, there is the curious parallel between the leptons and quarks in
each generation.  Each generation has a lepton with mass, a massless or
almost massless neutrino, and two quarks.  The massive lepton has charge
-1, the neutrino has charge 0 as its name suggests, the "down" type
quark has charge -1/3, and the "up" type quark has charge 2/3.  Funny
pattern, eh?  The Standard Model does not really explain this, although
it would be ruined by "anomalies" --- certain nightmarish problems that
can beset a quantum field theory --- if one idly tried to mess with the
generations by leaving out a quark or the like.  Indeed, this is why the
charm quark was first predicted, before the generation pattern was fully
apparent.  It's natural to try to "unify" the quarks and leptons, and
indeed, in grand unified theories like the SU(5) theory proposed in 1974
of Georgi and Glashow, the quarks and leptons are treated in a unified
way.
Another interesting pattern is the repetition of generations itself.  
Why is there more than one?  Why are there three, almost the same,
but with the masses increasing dramatically as we go up?   The Standard
Model makes no attempt to explain this, although it does suggest that
there had better be more than 17 quarks --- more, and the strong force
would not be "asymptotically free" (weak at high energies), which would
cause lots of problems for the theory.  In fact, experiments strongly
suggest that there are no more than 3 generations.  Why?  
Finally, there is the grand distinction between bosons and fermions.  
What does this mean?  Here we understand quite a bit from basic
principles.  For example, the "spin-statistics theorem" explains why
particles with half-integer spin should satisfy the Pauli exclusion
principle, while those with integer spin should like to hang out
together.  This is a very beautiful result with a deep connection to
topology, which I try to explain in 
2) John Baez, Spin, statistics, CPT and all that jazz, 
http://math.ucr.edu/home/baez/spin.stat.html
But many people have tried to bridge the chasm between bosons and
fermions, unifying them by a principle called "supersymmetry".  As in
the other cases mentioned above, when they do this, they then need to
pull tricks to "break" the symmetry to get a theory that fits the
experimental fact that bosons and fermions are very different.
Personally, I'm suspicious of all these symmetries postulated only to be
cleverly broken; this approach was so successful in dealing with the
electroweak force --- modulo the missing Higgs! --- that it seems to
have been accepted as a universal method of having ones cake and eating
it too.  
Now, string theory comes in two basic flavors.  Purely bosonic
string theory lives in 26 dimensions and doesn't have any fermions in
it.  Supersymmetric string theories live in 10 dimensions and have both
bosons and fermions, unified via supersymmetry.  To deal with the 
fermions in nature, most work in physics has focused on the
supersymmetric case.  Just for completeness, I should point out that
there are 5 different supersymmetric string theories: type I, type
IIA, type IIB, E8 x E8 heterotic and SO(32) heterotic.  For more on
these, see "week72".  We won't be getting into them here.  Instead,
I just want to explain how fermions work in different dimensions, and
why nice things happen in dimensions of the form 8k + 2.  Most of
what I say is in Section 3 of
3) John H. Schwarz, Introduction to supersymmetry, in Superstrings
and Supergravity, Proc. of the 28th Scottish Universities Summer
School in Physics, ed. A. T. Davies and D. G. Sutherland, University
Printing House, Oxford, 1985.
but mathematicians may also want to supplement this with material
from the book "Spin Geometry" by Lawson and Michelson, cited in
"week82".  
To understand fermions in different dimensions we need to understand
Clifford algebras.  As far as I know, when Clifford originally invented 
these algebras in the late 1800s, he was trying to generalize Hamilton's
quaternion algebra by considering algebras that had lots of different
anticommuting square roots of -1.  In other words, he considered
an associative algebra generated by a bunch of guys e_1,...,e_n,
satisfying
e_i^2 = -1
for all i, and 
e_i e_j = - e_j e_i
whenever i is not equal to j.  I discussed these algebras in "week82"
and I said what they all were --- they all have nice descriptions in terms
of the reals, the complexes, and the quaternions.  
These original Clifford algebras are great for studying rotations in 
n-dimensional Euclidean space --- please take my word for this for now.
However, here we want to study rotations and Lorentz transformations
in n-dimensional Minkowski spacetime, so we need to work with a slightly 
Different kind of Clifford algebra, which was probably invented by Dirac.  
In n-dimensional Euclidean space the metric (used for measuring distances) 
is
                        dx_1^2 + dx_2^2 + ... + dx_n^2  
while in n-dimensional Minkowski spacetime it is
                       dx_1^2 + dx_2^2 + ... - dx_n^2
or if you prefer (it's just a matter of convention), you can
take it to be
                     - dx_1^2 - dx_2^2 - ... + dx_n^2
So it turns out that we need to switch some signs in the definition 
of the Clifford algebra when working in Minkowski spacetime.  
In general, we can define the Clifford algebra C_{p,q} to be the algebra 
generated by a bunch of elements e_i, with p of them being square roots 
of -1 and q of them being square roots of 1.  As before, we require that 
they anticommute:
e_i e_j = - e_j e_i
when i and j are different.  Physicists usually call these guys "gamma
matrices".  For n-dimensional Minkowski space we can work either
with C_{n-1,1} or C_{1,n-1}, depending on our preference.  As Cecile 
DeWitt has pointed out, it *does* make a difference which one we use.  
With some work, one can check that these algebras go like this:
C_{0,1}   R + R               C_{1,0}   C
C_{1,1}   R(2)                C_{1,1}   R(2)
C_{2,1}   C(2)                C_{1,2}   R(2) + R(2)
C_{3,1}   H(2)                C_{1,3}   R(4)
C_{4,1}   H(2) + H(2)         C_{1,4}   C(4)
C_{5,1}   H(4)                C_{1,5}   H(4)
C_{6,1}   C(8)                C_{1,6}   H(4) + H(4) 
C_{7,1}   R(16)               C_{1,7}   H(8)
I've only listed these up to 8-dimensional Minkowski spacetime, and
the cool thing is that after that they sort of repeat --- more precisely,
C_{n+8,1} is just the same as 16 x 16 matrices with entries in C_{n,1},
and C_{1,n+8} is just 16 x 16 matrices with entries in C_{n,1}!  
This "period-8" phenomenon, sometimes called Bott periodicity, has 
implications for all sorts of branches of math and physics.  This is
why fermions in 2 dimensions are a bit like fermions in 10 dimensions
and 18 dimensions and 26 dimensions....
In physics, we describe fermions using "spinors", but there are
different kinds of spinors: Dirac spinors, Weyl spinors, Majorana
spinors, and even Majorana-Weyl spinors.  This is a bit technical but
I want to dig into it here, since it explains what's special about
8k + 2 dimensions and especially 10 dimensions.  
Before I get technical, though, let me just summarize the point for 
those of you who don't want all the gory details.   "Dirac spinors"
are what you use to describe spin-1/2 particles that come in both
left-handed and right-handed forms and aren't their own antiparticle 
--- like the electron.  Weyl spinors have half as many components,
and describe spin-1/2 particles with an intrinsic handedness that 
aren't their own antiparticle --- like the neutrino.   "Weyl spinors"
are only possible in even dimensions!
Both these sorts of spinors are "complex" --- they have complex-valued 
components.  But there are also real spinors.  These are used for describing 
particles that are their own antiparticle, because the operation of 
turning a particle into an antiparticle is described mathematically
by complex conjugation.  "Majorana spinors" describe spin-1/2 particles 
that come in both left-handed and right-handed forms and are their 
own antiparticle.  Finally, "Majorana-Weyl spinors" are used to describe 
spin-1/2 particles with an intrinsic handedness that are their own
antiparticle.  
As far as we can tell, none of the particles we've seen are Majorana
or Majorana-Weyl spinors, although if the neutrino has a mass it
might be a Majorana spinor.  Majorana and Majorana-Weyl spinors
only exist in certain dimensions.  In particular, Majorana-Weyl spinors
are very finicky: they only work in dimensions of the form 8k + 2.  
This is part of what makes supersymmetric string theory work in 10 
dimensions!
Now let me describe the technical details.  I'm doing this mainly
for my own benefit; if I write this up, I'll be able to refer to
it whenever I forget it.  For those of you who stick with me, there
will be a little reward: we'll see that a certain kind of supersymmetric 
gauge theory, in which there's a symmetry between gauge bosons and 
fermions, only works in dimensions 3, 4, 6, and 10.  Perhaps 
coincidentally --- I don't understand this stuff well enough to know ---
these are also the dimensions when supersymmetric string theory works
classically.  (It's the quantum version that only works in dimension 10.)
So: part of the point of these Clifford algebras is that they give 
representations of the double cover of the Lorentz group in different
dimensions.  In "week61" I explained this double cover business,
and how the group SO(n) of rotations of n-dimensional Euclidean space 
has a double cover called Spin(n).  Similarly, the Lorentz group
of n-dimensional Minkowski space, written SO(n-1,1), has a double cover 
we could call Spin(n-1,1).  The spinors we'll discuss are all 
representations of this group.  
The way Clifford algebras help is that there is a nice way to
embed Spin(n-1,1) in either C_{n-1,1} or C_{1,n-1}, so any 
representation of these Clifford algebras gives a representation
of Spin(n-1,1).   We have a choice of dealing with real representations or 
complex representations.  Any complex representation of one of
these Clifford algebras is also a representation of the *complexified* 
Clifford algebra.   What I mean is this: above I implicitly wanted
C_{p,q} to consist of all *real* linear combinations of products of 
the e_i, but we could have worked with *complex* linear combinations 
instead.  Then we would have "complexified" C_{p,q}.  Since the
complex numbers include a square root of minus 1, the complexification
of C_{p,q} only depends on the dimension p + q, not on how many minus 
signs we have. 
Now, it is easy and fun and important to check that if you complexify R 
you get C, and if you complexify C you get C + C, and if you complexify 
H you get C(2).  Thus from the above table we get this table: 
dimension n        complexified Clifford algebra
    1                  C + C
    2                  C(2)
    3                  C(2) + C(2)
    4                  C(4)
    5                  C(4) + C(4)
    6                  C(8)
    7                  C(8) + C(8)
    8                  C(16)
Notice this table is a lot simpler --- complex Clifford algebras
are "period-2" instead of period-8.  
Now the smallest complex representation of the complexified Clifford
algebra in dimension n is what we call a "Dirac spinor".  We can figure
out what this is using the above table, since the smallest complex 
representation of C(n) or C(n) + C(n) is on the n-dimensional complex
vector space C^n, given by matrix multiplication.  Of course, for 
C(n) + C(n) there are *two* representations depending on which copy 
of C(n) we use, but these give equivalent representations of Spin(n-1,1), 
which is what we're really interested in, so we still speak of "the" 
Dirac spinors.
So we get:
dimension n       Dirac spinors 
     1                 C
     2                 C^2
     3                 C^2
     4                 C^4
     5                 C^4
     6                 C^8
     7                 C^8
     8                 C^16
The dimension of the Dirac spinors doubles as we go to each new
even dimension.
We can also look for the smallest real representation of C_{n-1,1}
or C_{1,n-1}.  This is easy to work out from our tables using
the fact that the algebra R has its smallest real representation 
on R, while for C it's on R^2 and for H it's on R^4.  
Sometimes this smallest real representation is secretly just the 
Dirac spinors *viewed as a real representation* --- we can view C^n
as the real vector space R^{2n}.   But sometimes the Dirac spinors 
are the *complexification* of the smallest real representation ---
for example, C^{2n} is the complexification of R^n.   In this
case folks call the smallest real representation "Majorana spinors". 
When we are looking for the smallest real representations, we get 
different answers for C_{n-1,1} and C_{1,n-1}.  Here is what we get:
n   C_{n-1,1}      smallest         C_{1,n-1}      smallest 
                   real rep                        real rep 
1    R + R        R     Majorana     C             R^2    
2    R(2)         R^2   Majorana     R(2)          R^2   Majorana
3    C(2)         R^4                R(2) + R(2)   R^2   Majorana
4    H(2)         R^8                R(4)          R^4   Majorana  
5    H(2) + H(2)  R^8                C(4)          R^8 
6    H(4)         R^16               H(4)          R^16  
7    C(8)         R^16               H(4) + H(4)   R^16  
8    R(16)        R^16  Majorana     H(8)          R^32
I've noted when the representations are Majorana spinors.  Everything
repeats with period 8 after this, in an obvious way.
Finally, sometimes there are "Weyl spinors" or "Majorana-Weyl"
spinors.  The point is that sometimes the Dirac spinors, or
Majorana spinors, are a *reducible* representation of Spin(1,n-1).
For Dirac spinors this happens in every even dimension, because the 
Clifford algebra element Gamma = e_1 ... e_n commutes with everything
in Spin(1,n-1) and Gamma^2 is 1 or -1, so we can break the space of 
Dirac spinors into the two eigenspaces of Gamma, which will be smaller 
reps of Spin(1,n-1) --- the "Weyl spinors".  Physicists usually call this 
Gamma thing "gamma_5", and it's an operator that represents parity 
transformations.  We get "Majorana-Weyl" spinors only when we have
Majorana spinors, n is even, and Gamma^2 = 1, since we are then working
with real numbers and -1 doesn't have a square root.  You can work out
Gamma^2 for either and C_{n-1,1} or C_{1,n-1}, and see that we'll
only get Majorana-Weyl spinors when n = 8k + 2.
Whew!  Let me summarize some of our results:
n    Dirac       Majorana       Weyl     Majorana-Weyl
1     C            R             
2     C^2          R^2           C           R
3     C^2          R^2 
4     C^4          R^4           C^2
5     C^4       
6     C^8                        C^4
7     C^8
8     C^16         R^16          C^8
When there are blanks here, the relevant sort of spinor doesn't
exist.  Here I'm not distinguishing Majorana spinors that come from
C_{n-1,1} and those that come from C_{1,n-1}; you can do that with
the previous table.  Again, things continue for larger n in an obvious
way.  
Now, let's imagine a theory that has a supersymmetry between a gauge
bosons and a fermion.  We'll assume there are as many physical degrees of 
freedom for the gauge boson as there are for the fermion.   Gauge
bosons have n - 2 physical degrees of freedom in n dimensions: for
example, in dimension 4 the photon has 2 degrees of freedom, the spin-up
and the spin-down states.  So we want to find a kind of spinor that
has n - 2 physical degrees of freedom.  But the number of physical
degrees of freedom of a spinor field is half the number of (real) 
components of the spinor, since the Dirac equation relates the
components.  So we are looking for a kind of spinor that has 2(n - 2)
real components.  This occurs in only 4 cases:
n = 3  ->  2(n-2) = 2, and Majorana or Weyl spinors have 2 real components
n = 4  ->  2(n-2) = 4, and Majorana or Weyl spinors have 4 real components
n = 6  ->  2(n-2) = 8, and Weyl spinors have 8 real components
n = 10 ->  2(n-2) = 16, and Majorana-Weyl spinors have 16 real components
Note we count complex components as two real components.  And note how
dimension 10 works: the dimension of the spinors grows pretty fast as
n increases, but the Majorana-Weyl condition reduces the dimension by
a factor of 4, so dimension 10 just squeaks by!
John Schwarz explains how nice things happen in the same dimensions
for superstring theory in:
4) John H. Schwarz, Introduction to superstrings, in Superstrings
and Supergravity, Proc. of the 28th Scottish Universities Summer
School in Physics, ed. A. T. Davies and D. G. Sutherland, University
Printing House, Oxford, 1985.
He also makes a tantalizing remark: perhaps these 4 cases correspond
somehow to the reals, complexes, quaternions and octonions.  Note:
3 = 1 + 2, 4 = 2 + 2, 6 = 4 + 2 and 10 = 8 + 2.  You can never tell
with this stuff... everything is related.
-----------------------------------------------------------------------
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 by anonymous ftp from math.ucr.edu; they are in the
subdirectory pub/baez.  The README file lists the contents of all the
papers.  On the World-Wide Web, you can get these files by going to
http://math.ucr.edu/home/baez/
A complete index of the old issues of "This Week's Finds" is available
at
http://math.ucr.edu/home/baez/twf.html
but if you are cursed with a slow connection and just want a jumping-off
place to the olds issues, go to
http://math.ucr.edu/home/baez/twfshort.html
For the latest issue, go to
http://math.ucr.edu/home/baez/this.week.html
Return to Top
Subject: Poincare series for the depths of roots in a root system
From: dima@win.tue.nl (Dmitrii V. Pasechnik)
Date: 28 Oct 1996 14:30:51 +0100
The depth of a positive root  a in a root system F related to a Coxeter group G
is the minimal length of g in G such that the root  g a  is negative.
(Equivalently, if the refelection r_a has length l then depth a = (l+1)/2)
We conjecture that the function P(t)=a_1 t + a_2 t^2 + a_3 t^3 +..., where
a_i is the number of roots of depth i,
is always rational.
Some (computational) evidence to support this.
First,
For affine Coxeter groups a_i is a periodic function (of i) with a_i determined
in a simple way from a_i's for the associated finite group.
For instance, for affine A_n all a_i=n.
(this is not (yet) a theorem, but we suspect it's not so difficult to show.)
Second,
I give a (conjectural) recurrence relation for each case, and in the 1st case I
also write down the generating function.
(The latter is just a straightforward computation given the
recurrence. One could also compute a(n) as a function of n, although
it wouldn't look too nice.)
Note that for the sequence to satisfy a recurrence relation is
equivalent to have a rational generating function, cf. e.g.
R.Stanley "Enumerative Combinatorics I", Wadsworth 1986, Chapter 4.
1) o--o	   n: 1   2   3    4    5    6    7    8    9   10   11 
   |\/|   2*( 2   3   6   12   27   60  138  315  726 1668 3843....)
   |/\|
   o--o      a(n)=2a(n-1)+2a(n-2)-3a(n-3).
	F(x)=\sum a(n) x^n=2x((2-x-10x^2)/(1-2x-2x^2+3x^3)).
2) o--o	      1   2   3    4    5    6    7    8    9   10   11 
   | /|       4   5   8   13   24   44   83  158  303  582 1120
   |/ |
   o--o      a(n)=2a(n-1)-a(n-5).
3) o--o	      1   2   3    4    5    6    7    8    9   10   11 
   | /	      4   4   5    6    8   11   15   21   30   43   62
   |/ 
   o--o      a(n)=2a(n-1)-a(n-2)+a(n-3)-a(n-4) (for n>5).
4)
o
|\
| o---o---o
|/
o
The depths are as follows:
[ 5, 5, 6, 8, 11, 16, 25, 38, 59, 93, 148, 235, 376, 602, 966, 1550, 2491, 
  4003, 6436, 10348, 16643, 26766, 43052, 69247 ]
The recurrence is as follows:
a(n+1)=\sum_{i=n-11}^n v(i)*a(i), for v=[0,0,-1,-1,-2,-1,0,2,2,1,0].
I would appreciate receiving any comments on this.
(please copy your reply to my email dima@win.tue.nl)
Dmitrii V. Pasechnik
Department of Mathematics
Eindhoven University of Technology
PO Box 513, 5600 MB Eindhoven
The Netherlands
e-mail: dima@win.tue.nl
http://www.can.nl/~pasec
Return to Top
Subject: Galois Field Problem
From: Amr Mohamed Youssef
Date: 28 Oct 1996 17:28:21 GMT
Hi every one,
I have the following problem:
Let
A=A(x) be an arbitrary polynomail over GF(2^n), 
P=x^m + b1 x^(m-1) + .....+bm,
    dP
P'=---
    dx
      def
r(A,P) = A+A^2 + A^4 + A^8+  .....+A^(2^n-1)    mod P
A P' =c1 x^(m-a) +.....+cm   mod P
then it can be shown that
r(A,P)=c1.
For example r(x,P)=-b1.
The above result was proved by Carlitz[1] in realation to some exponential sum
problems.
My problem now is to calcualte r( (X+a)^-1 , P).
Thanks in advance.
[1]L. Carlitz, On certain functions connected with polynomials in a Galois filed,
Duke math. journal, Vol 1, 1035, pp.137-168
Return to Top
Subject: Re: surface and curve intersection algorithm
From: Saul Youssef
Date: Mon, 28 Oct 1996 12:52:15 -0500
heltne@@uu4.psi.com wrote:
> 
> In , Jacques Leroy  writes:
> >Desperately seeking any infos (source code, references, articles, ...)
> >about
> >surface and curve intersection algorithms.
> 
 You might try: 
   Multi-dimensional Searching and Computational Geometry, by
      Kurt Mehlhorn
   Algorithms in Combinatorial Geometry, by Herbert Edelsbrunner.
 Both are published by Springer-Verlag.
Saul Youssef
Return to Top
Subject: Re: diffeomorphism
From: Pat.McSwiggen@UC.Edu (Patrick D. McSwiggen)
Date: Mon, 28 Oct 1996 14:24:47 -0500
In article <54sln6$bms@mikasa.iol.it>, pcascini@hotmail.com (paolo cascini)
wrote:
>I'm not English so you have to soory my language...
>
>can you find two space that are diffeomorph of class Cn for each n, but 
>that are not diffepmorph of class c-inf.
No. See Hirsch, "Differential Topology", p. 50, Th 2.7.
If two C^s manifolds are C^1 diffeomorphic, then they are C^s
diffeomorphic. Here s can be infinity.
The interesting examples come from manifolds that are homeomorphic (only).
-- 
Patrick D. McSwiggen              Internet:   pat.mcswiggen@uc.edu
Department of Mathematics                  
University of Cincinnati
Return to Top
Subject: Re: This Week's Finds in Mathematical Physics (Week 93)
From: baez@math.ucr.edu (john baez)
Date: 28 Oct 1996 12:49:31 -0800
Joshua Burton and others have caught some errors.  Perhaps the
only direct relevance to mathematics is the prehistory of anomalies
(see below), but I don't want to be telling mathematicians a bunch
of bogus physics.
In article <55153l$r58@charity.ucr.edu>, John Baez  wrote:
>    ELECTROMAGNETIC FORCE          WEAK FORCE         STRONG FORCE
>        
>    photon                          W+                  6 gluons 
>                                    W-
>                                    Z  
There are 8 gluons.  
>Right now, we believe that neutrinos only exist in a
>left-handed form, rotating one way but not the other around the axis
>they move along.  This is intimately related to their apparent
>masslessness.  In fact, for reasons that would take a while to explain,
>the lack of parity symmetry in the Standard Model forces us to assume
>all fermions acquire their mass only through interaction with the Higgs
>particle!  
Joshua notes that were there a right-handed neutrino, it would be
an SU(3) x SU(2) x U(1) singlet (or in math lingo: it'd transform
in the trivial representation), so it could acquire a Majorana mass.
One proposed method is the "seesaw mechanism".  For more on this sort
of thing see the recent thread on sci.physics.research entitled 
"Proof of neutrino rest mass?" and also 
Paul Langacker, Implications of neutrino mass,
http://dept.physics.upenn.edu/~www/neutrino/jhu/jhu.html 
>The Standard Model [...]
>would be ruined by "anomalies" --- certain nightmarish problems that
>can beset a quantum field theory --- if one idly tried to mess with the
>generations by leaving out a quark or the like.  Indeed, this is why the
>charm quark was first predicted, before the generation pattern was fully
>apparent.  
This is wrong.  Joshua notes that while the famous 1969 Adler-Bell-Jackiw
paper on anomalies predates the discovery of the charm quark, and perhaps 
*could* have been used to predict the charm quark, the actual prediction 
of the charm quark by Glashow-Iliopoulos-Maiani in 1970 was made for 
other reasons.  
They proposed this quark, and the so-called "GIM mechanism", to solve some
problems concerning kaon decay.  In the Standard Model as it then stood,
a neutral K-bar meson could decay into a muon-antimuon pair like this:
             W+
dbar--<----~~~~~~~----<---mu+ 
           |     |
          u^     V nu
           |     |
   s-->----~~~~~~~---->---mu-
            W-
(The neutral K-bar is a strange-antidown pair.)   But this decay was
not observed.  Glashow Iliopoulos and Maiani posited the new c quark
so that the process would be cancelled, or almost cancelled, by 
             W+
dbar--<----~~~~~~~----<---mu+
           |     |
          c^     V nu
           |     |
   s-->----~~~~~~~---->---mu-
            W-
Actually Joshua spoke of "neutral kaon mixing" rather than kaon decay,
and I'm getting the above stuff from Kerson Huang's book "Quarks Leptons
and Gauge Fields".  Perhaps the same general problem shows up in something
like the following? 
            W+
dbar--<----~~~~~~~----<---sbar
           |     |
          u^     V u 
           |     |
   s-->----~~~~~~~---->---d
            W-
In any event, Joshua says that his advisor, Mary Gaillard, noted that
the cancellation wouldn't be complete, since the up and charm have
different mass, and used the observed left-over kaon mixing to successfully
predict the charm mass six months before the charm quark --- or more
precisely, the J/psi, which is a c-cbar bound state --- was discovered
by Richter and Ting in 1974.  Ah, the good old days!  I was probably
learning all about F = ma at the time.
Later, this become touted as one of the reasons why one should
take anomalies seriously: without the charm quark to complete the
second generation, there would be a "triangle anomaly".   And by the time
I learned quantum field theory, "anomaly cancellation" had been established
as a religion in particle physics, to the great pleasure of mathematicians
such as those under whom I studied, for it served as a practical application
of the index theorem, which they were busy generalizing beyond all measure.
>The Standard
>Model makes no attempt to explain this, although it does suggest that
>there had better be more than 17 quarks --- more, and the strong force
>would not be "asymptotically free" (weak at high energies), which would
>cause lots of problems for the theory. 
There had better NOT be more than 17 of *something*.  Joshua says
there had better not be more than 17 generations, but Huang's book
says there had better not be more than 17 flavors of quark.  In any
event, we are pretty sure there are no more than 3 generations, unless
the next generation has a darn heavy neutrino or something else really
weird like that happens.
>Gauge
>bosons have n - 2 physical degrees of freedom in n dimensions: for
>example, in dimension 4 the photon has 2 degrees of freedom, the spin-up
>and the spin-down states.  
It would be better to say, "the left and right polarized states".
A corrected version is now at
http://math.ucr.edu/home/baez/week93.html
Return to Top
Subject: Re: Interesting DE Problem!!!
From: baez@math.ucr.edu (john baez)
Date: 28 Oct 1996 12:37:28 -0800
In article <54vvrn$h68@fsuj19.rz.uni-jena.de>,
Achim Flammenkamp  wrote:
>In article <54f3e4$gh6@amanda.dorsai.org>,
>Abe Mantell  wrote:
>>I have been quite a bit baffled by the following initial value problem:
>>y' = y^2 + x^2, y(0)=1
>There are different kinds of singularities (in one independent variable). 
>1) risable singularities
>1) These are trivial: like y=(x^2-1)/(x-1) has a singularity at x=1
In English these are usually called "removable" singularities, but
I sort of like the term "risible" singularities, meaning "laughable"
singularities.
However, the singularity displayed by y(t) satisfying the above equation
is no joke.
Return to Top
Subject: Re: terminology
From: ikastan@alumnae.caltech.edu (Ilias Kastanas)
Date: 29 Oct 1996 00:21:50 GMT
In article <32717FEF.6127@mailbox.syr.edu>,
Ray DeCampo   wrote:
>Michael Hardy wrote:
>> 
>>         Last week I wrote:
>> 
>> >         When, as in defining quotient groups, one defines an operation on
>> > classes by picking a member of each class to be operated on, and performing
>> > an operation on the members, (and then worrying about "problematic", (i.e.
>> > possibly ambiguous and possibly not) definitions), then one is defining an
>> > operation *******-wise.  I'm looking for this locution, filling in the
>> > row of asterisks with a one- or two-syllable word.  Does anyone know it?
>> 
>>         In a post that hasn't yet appeared at this site
>>         (I got it by gopher-ing to math.lfc.edu
>>         Andrew Stacey  replied:
>> 
>> > This is a complete stab in the dark, it's probably not recognised
>> > terminology but if you just want something that carries the idea try:
>> >
>> > representative-wise
>> >
>> > Since the 'x' in x+H is a representative element of the equivalence class
>> > defined by: x~x' iff x-x' is in H (using addition rather than mult as
>> > subtraction is easier to type than inverses!)
>> [cut critique of 'representative-wise']
>
>How about dropping the whole '-wise' suffix?  How about saying that the 
>operation is defined via representatives?  (or 'on representatives' or 
>{your favorite preposition here} representatives?)
	How about  "by proxy"?
					Ilias
Return to Top
Subject: infinitely near points
From: adler@pulsar.cs.WKU.EDU (Allen Adler)
Date: 28 Oct 1996 23:07:13 -0600
Let X denote the complex projective plane. Let x0 be a point
of X. Blow up x0 and choose a point x1 on the resulting exceptional
line. Blow up x1 and choose a point x2 on the resulting exceptional
line and so on. Stop when you have chosen xn.
Can you describe the orbits of the collineation group G of X on the
variety of all possible choices of (x0,...,xn)? 
Allan Adler
adler@pulsar.cs.wku.edu
Return to Top

Downloaded by WWW Programs
Byron Palmer