From: kramsay@aol.commangled (Keith Ramsay)
Subject: Re: Query about Category Theory.
Date: 19 Jun 2000 03:33:25 GMT
Newsgroups: sci.math
Summary: [missing]
In article <8ihn4u$6uf$2@cantuc.canterbury.ac.nz>,
mathwft@math.canterbury.ac.nz (Bill Taylor) writes:
|Does anyone know of any cases where Category Theory (morphisms, functors &
|all that) has helped solve an unsolved problem?
I believe it helped Grothendieck and Deligne arrive at some of their
ideas, like the ones used in Deligne's proof of the Weil conjecture.
Everything done using �tale cohomology has a bit of categorical
machinery running in the background.
It helped me with a result on a family of elliptic curves (my half of
a joint paper with Rajiv Gupta at UBC). As I worked on the problem, I
realized that various of my constructions were actually simply
pullbacks. There was a group action as well.
Suppose you have a family E_t of elliptic curves, parametrized by t.
In our key cases, the curves E_t all had extra automorphisms. If you
started with some E_t and "followed it around" as t circled one of the
points where the family degenerates, it might come back with an
automorphism having been applied to it. Then there's a set of points
on each one, and following the points as t varies, you get a (set of)
curves (complex curves, i.e. Riemann surfaces). The automorphism
causes some points to come around so that they're on the same
component as one of the "other" points. There are in general fewer
components than points being tracked; these points are sections of
curves which may have multiple branches over the base (t). It was
important to know the irreducible components of this collection of
curves. I'd done some concrete calculations, but I was still finding
it easy to slip up.
Considered in concrete terms, I was finding it confusing to see how
the group action played itself out, but a categorical lemma about what
happens when you have a pullback diagram with a group acting on all of
the objects (in a compatible way) helped out. It was a case where the
usual category theorists' claim of having a useful higher level of
abstraction proved to be true. This one fact was easiest for me to
understand as a general fact about this situation, rather than as some
observation about how these Riemann surfaces wrap around the branch
points.
Functors between categories "& all that" didn't play an obvious role,
although I think it was of psychological help to know that (most of)
what I was doing could be placed in the birational category, where up
to birational equivalence the singularities aren't a problem (algebraic
varieties are birationally equivalent to nonsingular ones). I suppose
some of it could be interpreted as relating the category of varieties
over the rationals with the category of varieties over the rationals
with a G action on them, and the group action can itself be thought of
as a functor.
This was the paper which was rejected at the first journal we sent it
to, on the advice of a referee who said we should state early on that
n was an integer. Paragraph 1 described the problem in general.
Paragraph 2 started by saying that throughout the paper, n would be
an integer. :-(
Keith Ramsay
==============================================================================
From: baez@galaxy.ucr.edu (John Baez)
Subject: Re: Query about Category Theory
Date: 21 Jun 2000 15:19:53 GMT
Newsgroups: sci.math
In article <8ihn4u$6uf$2@cantuc.canterbury.ac.nz>,
Bill Taylor wrote:
>Does anyone know of any cases where Category Theory (morphisms, functors & all
>that) has helped solve an unsolved problem?
>
>That is, a problem in some other branch of math, posed without reference to
>categorical ideas, and previously unsolved, that was first solved via
>Category T.
>
>I realize that CT provides a unifying framework for many seemingly disparate
>ideas in math, and that is a fine thing of course; but I was just wondering
>if it had this problem-solving capability.
>
>TIA for any helpful responses.
This question about categories was forwarded to the category theory
mailing list ... I append a reply. Briefly, the answer is that
category theory has been used for solving *many* problems. A
famous early example is the use of topos theory by Grothendieck
and Deligne for solving the Weil conjectures... but by now abelian
categories and derived categories are part of the working toolkit of
anyone in homological algebra, and model categories are similarly part
of the homotopy theorist's repertoire... because they help solve concrete
problems! In physics, the rigorous construction of 3d TQFTs using
monoidal categories is a nice example.
.........................................................................
Date: Wed, 21 Jun 2000 15:23:33 +0100
From: Ronnie Brown
To: categories@mta.ca
Subject: categories: Re: A question on:
I have a paper with Loday on calculating the third homotopy group of a
suspension of a K(G,1)
51. (with J.-L. LODAY), ``Van Kampen theorems for diagrams of spaces'',
{\em Topology} 26
(1987) 311-334.
in terms of a new tensor product of non abelian groups. It would not be a
sensible task to rewrite the paper without category theory (considered as a
unifying principle, as a mode for efficient calculation in certain
algebraic structures, and as a supplier of new algebraic structures).
To go further back (and higher, of course) Grothendieck's extension of the
Riemann-Roch Theorem uses category theory explicitly. In the 1950's people
were trying to give algebraic proofs of this theorem, when AG came up with
an algebraic proof of a vast generalisation. Then there is the categorical
background to the proof of the Weil conjectures ....
But the question is misplaced - in the early part of the 20th century, I
expect some would ask if set theory was really necessary for a confirmed
problem solver! The history of maths shows that maths greatest contribution
to science, culture and technology has been in terms of expressive power,
to give a language for intuitions which enables exact description,
calculation, deduction. (Exam question: discuss the last statement, with an
emphasis on particular examples!) It also allows for the *formulation* of
new problems, which perhaps cause old interests to lapse as people perceive
there are more exciting things to do. It is a narrow view to regard
`important maths' as necessarily that which solves well known problems, and
so leave evaluation as akin to a sports league table: how old is the
problem? who has worked on it? etc, etc. The progress of maths is much more
complicated and interesting than that! It won't help the public image of
maths if it is seen that mathematicians believe the most important aspect
of their subject is a (to the public) bizarre set of problems which seem to
interest no one else.
However, it is important that these questions be asked, together with
questions on modes for evaluating `good maths'. As AG remarked in a letter,
maths was held back for centuries for lack of the `trivial' concept of
zero! For more discussion, look at
http://www.bangor.ac.uk/ma/CPM/cdbooklet/knots-m.html
knots2.html
So let the debate be broadened and continued!
Ronnie Brown
(will someone please forward this to the newsgroup?)
PS I am trying to trace a combinatorics problem solved in the 1950s using
categories of paths, and which was at the time held up as the sort category
theory could not do! Maybe my memory is failing! But the question put on
the newsgroup is an old war horse!
-------- Original Message --------
Subject: categories: A question on:
Date: Mon, 19 Jun 2000 17:13:05 +1000
From: Ross Street
To: "categories@mta.ca"
I thought the CatNet would be interested in the following question
which appeared on . The letter was pointed out
to me by a colleague at Macquarie.
--Ross
*************************************************
From: mathwft@math.canterbury.ac.nz (Bill Taylor)
Newsgroups: sci.math
Subject: Query about Category Theory.
Date: 18 Jun 2000 05:36:30 GMT
Organization: Department of Mathematics and Statistics, University of
Canterbury, Christchurch, NewZealand
Does anyone know of any cases where Category Theory (morphisms, functors &
all that) has helped solve an unsolved problem?
That is, a problem in some other branch of math, posed without reference to
categorical ideas, and previously unsolved, that was first solved via
Category T.
I realize that CT provides a unifying framework for many seemingly disparate
ideas in math, and that is a fine thing of course; but I was just wondering
if it had this problem-solving capability.
TIA for any helpful responses.
----------------------------------------------------------------------------
--- Bill Taylor W.Taylor@math.canterbury.ac.nz