From: Erik Palmgren <palmgren@math.uu.se>
To: Thomas Streicher <streicher@mathematik.tu-darmstadt.de>
Cc: Toby Bartels <toby+categories@ugcs.caltech.edu>, categories@mta.ca
Subject: Re: Is equality evil?
Date: Sun, 19 Sep 2010 22:30:31 +0200 (CEST) [thread overview]
Message-ID: <E1Oximi-0005OL-Jc@mlist.mta.ca> (raw)
In-Reply-To: <E1OxKoj-0004xW-B9@mlist.mta.ca>
Thanks Thomas for bringing up the intensional type theory perspective. Here
equality of objects in categories is a serious problem already for sets,
or rather for their type theoretic counterpart, setoids (types with
equivalence classes). The so-called identity types give a minimal
equality on each type A: Id(A,a,b) are the proofs that a and b
are equal in A. But the fact that these proofs are in general not
unique and induces a groupoid structure makes it difficult to apply
when e.g. A is a universe of types. Some remarks are contained in
in this preprint
http://www.mittag-leffler.se/preprints/files/IML-0910f-36.pdf
If one tries to use the Id-types to define equality on (small) setoids
one ends up with a groupoid of objects rather than a setoid objects.
I think the situation can be described as follows for a general category K,
which we think of as a category without equality of objects. We now
want to induce an equality of objects on (a part of) K. Take a
groupoid G and a functor F:G -> K. We define the groupoids of
objects C_0, arrows C_1 and composable arrows C_2 in a natural
way. C_0 is G and C_1 has for objects triples (a,b,f) where a,b in G
and f:F(a) -> F(b), and a morphism (a,b,f) -> (a',b',f') is a pair of
G morphisms r:a -> a' and s:b -> b' with F(s) f = f' F(r). C_2 has for
objects pairs of morphisms composable with an mediating map m from G,
i.e. ((a,b,f),(c,d,g),m). The composition is g F(m)f. The morphisms of
C_2 are then pairs of C_1 morphisms
((r,s), (t,u)): ((a,b,f),(c,d,g),m) -> ((a',b',f'),(c',d',g'),m')
with m's= t m.
When G is a discrete groupoid this becomes a standard category. When G is
a groupoid it becomes almost a category. Notions of limits can be
formulated in a natural fashion. At least when K=Setoids, C inherits
limits from K. Perhaps some general results are known about this
construction of C from functor F:G -> K on a groupoid G?
Erik
> It's no problem to come up with logics without equality predicates. Just
> omit the equality predicate and its associated axioms. For ideological reasons
> most logic texts give equality a distinguished status for historical reasons.
> In case of extensional theories it suffices to have equality on base type and
> lift it `a la logical relations. But then there might be operations which
> don't respect this kind of equality. In other words identity types are not
> necessary for doing constructive mathematics.
> Intensional Id-types arise from putting the idea to an extreme that all logical
> concepts are "inductively defined", i.e. are given by constructors and
> eliminators.
> The notion of equality of types you refer to is a different one. Namely
> judgemental equality which cannot be used as an ingredient for forming
> propositions.
>
......
[For admin and other information see: http://www.mta.ca/~cat-dist/ ]
next prev parent reply other threads:[~2010-09-19 20:30 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-09-15 11:43 are fibrations evil? Thomas Streicher
2010-09-16 0:28 ` Michael Shulman
2010-09-16 1:14 ` Peter LeFanu Lumsdaine
2010-09-16 5:14 ` Is equality evil? Vaughan Pratt
2010-09-17 8:28 ` Toby Bartels
2010-09-18 14:11 ` Thomas Streicher
2010-09-19 20:30 ` Erik Palmgren [this message]
2010-09-24 12:50 ` Bas Spitters
[not found] ` <20100918141110.GC9467@mathematik.tu-darmstadt.de>
2010-09-22 4:00 ` Toby Bartels
2010-09-25 16:18 ` Michael Shulman
[not found] ` <20100922040041.GB14958@ugcs.caltech.edu>
2010-09-22 10:27 ` Thomas Streicher
2010-09-16 8:50 ` why it matters that fibrations are "evil" Thomas Streicher
[not found] ` <AANLkTinosTZ2NQW9biPxiwpX9zPi5m=kwvA16nHjK=Xu@mail.gmail.com>
2010-09-16 9:47 ` are fibrations evil? Thomas Streicher
2010-09-16 10:00 ` Prof. Peter Johnstone
[not found] ` <alpine.LRH.2.00.1009161023190.12162@siskin.dpmms.cam.ac.uk>
2010-09-16 10:46 ` Thomas Streicher
2010-09-17 7:44 ` Toby Bartels
[not found] ` <20100916094755.GA19976@mathematik.tu-darmstadt.de>
2010-09-17 5:01 ` Michael Shulman
2010-09-18 13:48 ` Thomas Streicher
[not found] ` <20100918134829.GB9467@mathematik.tu-darmstadt.de>
2010-09-20 16:25 ` Michael Shulman
2010-09-24 15:30 Is equality evil? Mattias Wikström
2010-09-25 0:16 Is equality evil? Fred E.J. Linton
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=E1Oximi-0005OL-Jc@mlist.mta.ca \
--to=palmgren@math.uu.se \
--cc=categories@mta.ca \
--cc=streicher@mathematik.tu-darmstadt.de \
--cc=toby+categories@ugcs.caltech.edu \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).