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).
    • CommentRowNumber1.
    • CommentAuthorUrs
    • CommentTimeJan 23rd 2013
    • (edited Jan 23rd 2013)

    at complete Boolean algebra I fixed the missing op{}^{op} and then added a pointer to Set – Properties – Opposite category and Boolean algebras.

    • CommentRowNumber2.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 23rd 2013

    I added a few words on algebraicity aspects of complete Boolean algebras (they form a variety of infinitary algebras, which however is not monadic over SetSet).

    Incidentally, looking over the constructive-mathematics remarks of CABA’s – do I recall correctly that they can also be described (classically) as completely distributive complete Boolean algebras, and that this gives an alternative constructively acceptable definition?

    • CommentRowNumber3.
    • CommentAuthorTobyBartels
    • CommentTimeJan 24th 2013

    I don't know if it's constructively true that Set opSet^op is equivalent to the category of completely distributive complete Boolean algebras, but if so, it's not through the power set functor, which essentially never lands in Boolean algebras at all.

    • CommentRowNumber4.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 24th 2013

    Well, good point. I guess my comment was directed more at replacing the term “atomic” with the term “completely distributive”, whose equationality would seem less prone to raising constructivist concerns than the notion of being atomic.

    Is there an infinitary Lawvere theory whose models in any Grothendieck topos EE are given by the monadic functor P:E opEP: E^{op} \to E? As you point out, the theory of completely distributive complete Boolean algebras won’t work. But it doesn’t seem very likely to me now that there is any Lawvere theory description that would work uniformly – or if there is, how would one describe it?

    • CommentRowNumber5.
    • CommentAuthorMike Shulman
    • CommentTimeJan 24th 2013

    It does seem unlikely!

    • CommentRowNumber6.
    • CommentAuthorZhen Lin
    • CommentTimeJan 25th 2013
    • (edited Jan 25th 2013)

    Obviously, one has to fix some cardinal constraints – otherwise whatever Lawvere theory you pick, you could just go one universe up and note that the theory has a model that is not complete in the bigger universe. (Is it true that every non-degenerate Grothendieck topos satisfies Giraud’s axioms only in the universe in which it is defined? Completeness pushes the size of a Grothendieck topos upwards, while local-smallness pushes the size downwards.)

    Anyway, whatever we do, it cannot work, because models of an infinitary Lawvere theory (of whatever cardinality) can be pushed forward along right adjoints. So consider the complete atomic boolean algebra 22 and the skyscraper sheaf functor x *:SetSh(X)x_* : \mathbf{Set} \to \mathbf{Sh}(X) for some point xx in a sufficiently non-trivial topological space XX. Then x *2x_* 2 cannot be the power object of anything in Sh(X)\mathbf{Sh}(X): indeed, if x *2=PAx_* 2 = P A, then for any non-empty open subset UU not containing xx, we must have A(U)=A(U) = \emptyset, and for any open subset VV containing xx, we must have A(V)A(V) \ne \emptyset; but then we would have a map A(X)A (X) \to \emptyset with A(X)A (X) \ne \emptyset, because certainly xXx \in X.

    • CommentRowNumber7.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 25th 2013

    Nice argument, Zhen!

    • CommentRowNumber8.
    • CommentAuthorMike Shulman
    • CommentTimeJan 25th 2013

    Is it true that every non-degenerate Grothendieck topos satisfies Giraud’s axioms only in the universe in which it is defined?

    The notion of “Grothendieck topos” is always relative to a universe. A UU-Grothendieck-topos is locally UU-small, has a UU-small generating set, has UU-small limits and colimits (including UU-small disjoint and stable coproducts, which is the non-finitary one of Giraud’s exactness axioms), and is equivalent to the category of UU-small sheaves on a UU-small site. If you’re asking whether a nondegenerate UU-Grothendieck-topos might happen to also be a UU'-Grothendieck-topos for some other UU', then I think the answer is no, not if UUU\in U', since a UU-Grothendieck-topos would then be a UU'-small category, so by Freyd’s theorem if it were also UU'-cocomplete it would be a poset, and the only topos that’s a poset is the degenerate one.

    • CommentRowNumber9.
    • CommentAuthorZhen Lin
    • CommentTimeJan 25th 2013

    Thanks!

    That argument works generically for (co)complete locally small categories. It feels like something that should have consequences for universe-ambiguity in category theory, but I’ll have to chew on that idea more.

    • CommentRowNumber10.
    • CommentAuthorTobyBartels
    • CommentTimeJan 25th 2013

    @Todd: Being atomic is not what causes the problems constructively; being Boolean is. It's slightly nontrivial to formulate being atomic in a constructive manner (see discussion at atom), but when you do it, then it is true that power sets are atomic. (It's also true that they're completely distributive, of course.)

    • CommentRowNumber11.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 25th 2013

    Thanks, Toby. I haven’t expressed my thoughts very well (and anyway they’re in a different direction from the point you’re making): the notion of completely distributive complete Boolean algebra is manifestly expressible via an infinitary Lawvere theory. The axioms of such a theory, since they do not involve essential uses of disjunction, existential quantification, or negation, do not involve any non-trivial reformulations or undergo bifurcations when we move to a constructive setting.

    • CommentRowNumber12.
    • CommentAuthorTobyBartels
    • CommentTimeJan 27th 2013

    That's true.

    • CommentRowNumber13.
    • CommentAuthorDmitri Pavlov
    • CommentTimeApr 21st 2019

    Stone duality for complete Boolean algebras.

    diff, v12, current

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)