Not signed in (Sign In)

Start a new discussion

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

  • Sign in using OpenID

Discussion Tag Cloud

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

Welcome to nForum
If you want to take part in these discussions either sign in now (if you have an account), apply for one now (if you don't).
  1. Hi all, I added an entry for Leibniz’s identity of indiscernibles.

    I thought it would be nice to include a discussion of the way in which univalence refines the identity of indiscernibles, but I am just beginning to learn about UF and was not sure what to say.

    In Awodey’s 2013 exposition of univalence he states:

    “Rather than viewing it as identifying equivalent objects, and thus collapsing distinct objects, it is more useful to regard it as expanding the notion of identity to that of equivalence. For mathematical purposes, this is the sharpest notion of identity available; the question whether two equivalent mathematical objects are “really” identical in some stronger, non-logical sense, is thus outside of mathematics.”

    Thus we can regard univalence as “loosening” the notion of identity in such a way that it validates Leibniz’s law.

    Certainly, I can just add something like this. However, it seems to me that a philosophical notion of equality that accords with intuition may profitably make a distinction between isomorphic structures when they are not “really” equal ontologically. Awodey is making a stance about “mathematical” rather than “real” identity, as is considered in (some formulations?) of Leibniz’s Law.

    Is it possible to think about the connection to Leibniz’s Law like this:

    Equivalence is akin to and entails a kind of “observational equivalence”/indiscernibility.

    Thus we have the following putative principles:

    • Indiscernibility of Identicals–intuitively true
    • Identity of indiscernibles–controversial
    • Indiscernibility is indiscernible from identity–intuitively true, and maintains some of the intent of identity of indiscernibles

    So can we think of Univalence as a refinement that says “Indiscernibility is indiscernible from identity, so we may as well treat indiscernibles as equal by transporting isomorphisms into identifications.”? Or is this not the way to think about this/do folks in UF have a stronger stance such as “real” equality being absurd?

    Best, Colin

    • CommentRowNumber2.
    • CommentAuthorMike Shulman
    • CommentTimeJun 1st 2014

    The way I think about it, in type theory, identity of indiscernibles is just true, essentially by definition of “identity” – that’s the induction principle for identity types. Possibly the issue is rather that the notion of “indiscernibility” refers to a notion of “property”, and the question is what sort of properties we consider. In mathematics, we generally only consider properties that are invariant under equivalence; thus equivalent objects are indiscernible, and hence identical.

    • CommentRowNumber3.
    • CommentAuthorDavid_Corfield
    • CommentTimeJun 1st 2014

    So one question is, outside of mathematics, what would lead one to employ properties which are not invariant under equivalence. It would be important to remember that just as it is right to distinguish non-identical isomorphic subgroups of a group, so one should distinguish ’isomorphic’ subregions of the universe.

    • CommentRowNumber4.
    • CommentAuthorMike Shulman
    • CommentTimeJun 1st 2014

    I’m not sure what “equivalence” means outside of mathematics.

    • CommentRowNumber5.
    • CommentAuthorDavid_Corfield
    • CommentTimeJun 2nd 2014

    That was rather loose on my part. I was thinking about an applied type theory where two entities are given as the interpretations of two equivalent terms.

    Metaphysicians often think up simple universes to test out their intuitions. For example,they imagine two physically identical spherical objects as the sole occupants of a space. Leaving aside the matter of whether this is physically possible, the worry is that no properties distinguish the objects and yet we don’t want to think of them as identical.

    • CommentRowNumber6.
    • CommentAuthorZhen Lin
    • CommentTimeJun 2nd 2014

    Unless position is not considered to be a property, I don’t see any reason for the two spheres to be indiscernible. This is the same argument that shows that a set can have many different one-element subsets.

    • CommentRowNumber7.
    • CommentAuthorDavid_Corfield
    • CommentTimeJun 2nd 2014

    Right, so one side says “I can choose a ball and say that it has the property of being in the place it occupies, while the other ball doesn’t occupy that place. Properties differ.” The other side worries that in effect, you’ve placed yourself in the universe as observer, or at least introduced something extraneous to the universe itself.

    And yes, mathematics, of course, is able to cope with entities with autoequivalences permuting non-identical subparts.

  2. Many thanks for the interesting responses!

    It seems to me that the issue of whether there is a sense in which differently-presented instances of the same structure are "actually" distinct and the issue of evil properties "outside of math" are of a similar species since both involve bringing in an ontological notion of identity.
    • CommentRowNumber9.
    • CommentAuthorTobyBartels
    • CommentTimeJun 2nd 2014

    Identity of indiscernibles in higher-order category theory means equivalence of objects that are discernible only by evil means.

  3. Without Univalence, such predicates are of course definable: Define two types of natural numbers, say, N1 and N2, and let the predicate be P=λx.x=N1P = \lambda x. x = N1.

    Arvid Marx

    diff, v11, current

    • CommentRowNumber11.
    • CommentAuthoratmacen
    • CommentTimeDec 21st 2019

    Arvid, but you still can’t prove ¬P(N2)\not P(N2). That is, P doesn’t discern N1 and N2. So N1 and N2 are no longer identical, but they’re still indiscernible. Removing axioms never makes things discernible.

    Although it’s probably way too late, I think the original discussion missed an important point: Equivalence is a positive, constructive statement, while indiscernibility is negative. If anything, univalence is “identity of interchangeables”, which seems intuitively weaker than identity of indiscernibles. Also, type theory’s identity elimination rule seems like interchangeability of identicals, which is intuitively stronger than indiscernibility of identicals.

    • CommentRowNumber12.
    • CommentAuthorMike Shulman
    • CommentTimeDec 22nd 2019

    I think it’s reasonable to interpret “indiscernible” in a positive way in a constructive context, just as the reasonable constructive interpretation of “nonempty” is “inhabited”.

    • CommentRowNumber13.
    • CommentAuthorMike Shulman
    • CommentTimeDec 22nd 2019

    I rolled back the last edit; as Matt said, it was correct before with “provably distinguishes”.

Add your comments
  • Please log in or leave your comment as a "guest post". If commenting as a "guest", please include your name in the message as a courtesy. Note: only certain categories allow guest posts.
  • To produce a hyperlink to an nLab entry, simply put double square brackets around its name, e.g. [[category]]. To use (La)TeX mathematics in your post, make sure Markdown+Itex is selected below and put your mathematics between dollar signs as usual. Only a subset of the usual TeX math commands are accepted: see here for a list.

  • (Help)