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

Site Tag Cloud

2-categories 2-category 2-category-theory abelian-categories adjoint algebra algebraic algebraic-geometry algebraic-topology analysis analytic-geometry arithmetic arithmetic-geometry bundles calculus categories category category-theory chern-weil-theory cohesion cohesive-homotopy-theory cohesive-homotopy-type-theory cohomology colimits combinatorics comma complex-geometry computable-mathematics computer-science constructive cosmology definitions deformation-theory descent diagrams differential differential-cohomology differential-equations differential-geometry differential-topology digraphs duality education elliptic-cohomology enriched fibration finite foundations functional-analysis functor galois-theory gauge-theory gebra geometric-quantization geometry graph graphs gravity grothendieck group-theory harmonic-analysis higher higher-algebra higher-category-theory higher-differential-geometry higher-geometry higher-lie-theory higher-topos-theory homological homological-algebra homotopy homotopy-theory homotopy-type-theory index-theory infinity integration integration-theory k-theory lie-theory limit limits linear linear-algebra locale localization logic manifolds mathematics measure-theory modal-logic model model-category-theory monads monoidal monoidal-category-theory morphism motives motivic-cohomology newpage noncommutative noncommutative-geometry number-theory of operads operator operator-algebra order-theory pasting philosophy physics pro-object probability probability-theory quantization quantum quantum-field quantum-field-theory quantum-mechanics quantum-physics quantum-theory question representation representation-theory riemannian-geometry scheme schemes set set-theory sheaf simplicial space spin-geometry stable-homotopy-theory string string-theory subobject superalgebra supergeometry svg symplectic-geometry synthetic-differential-geometry terminology theory topology topos topos-theory type type-theory universal variational-calculus

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.
    • CommentAuthorMike Shulman
    • CommentTimeDec 8th 2016

    I added the definition of uniformly/almost located to located subspace, as well as attempted proofs that covert and compact subspaces of uniformly regular uniform spaces are uniformly located. Unfortunately covert sets and compact spaces — as opposed to compact locales — seem to be hard to come by constructively. It would be nice to have some more concrete examples. I also don’t know what to write for the “idea” section, nor what Toby had in mind for “topologically located”.

    • CommentRowNumber2.
    • CommentAuthorTobyBartels
    • CommentTimeDec 9th 2016

    I found the definition of topological locatedness thanks to Google Books. It has a similar flavour to the uniform case. Also an Idea section.

    • CommentRowNumber3.
    • CommentAuthorMike Shulman
    • CommentTimeDec 9th 2016

    Thanks! Topological locatedness seems very weak; in particular, singletons being topologically located is weaker than quasiregularity.

    I’d like to see more examples of located and non-located subsets, as well as examples of what located subspaces are used for. Where should I be looking?

    • CommentRowNumber4.
    • CommentAuthorTobyBartels
    • CommentTimeDec 9th 2016

    Well, Bishop uses metrically located subsets. A lot of hereditary properties fail in general but work for located subspaces: total boundedness comes to mind. Besides that, Section 7.3 of Troelstra & van Dalen seems to be readable on Google Books.

    • CommentRowNumber5.
    • CommentAuthorMike Shulman
    • CommentTimeDec 9th 2016
    • (edited Dec 9th 2016)

    I’m no longer convinced that Bridges’ “almost located” is the most natural generalization of metric locatedness to uniform spaces. If we translate almost-locatedness into a metric space, it says that for any b>0b\gt 0 there exists an a>0a\gt 0 such that for any xx, either yS,d(x,y)>a\forall y\in S, d(x,y)\gt a or yS,d(x,y)<b\exists y\in S, d(x,y)\lt b. But here aa could be much smaller than bb, whereas metric locatedness says that for any b>0b\gt 0 and any a<ba\lt b, the same disjunction holds for all xx.

    It seems to me that a more natural notion of uniform locatedness would be that for any entourages U,V,WU,V,W such that WVUW\circ V \subseteq U we have ¬V[S]U[S]=X\neg V[S] \cup U[S] = X. Whereas Bridges’ condition is implied by metric locatedness, I believe this condition implies metric locatedness, taking U=U bU=U_b and V=U a+ε V = U_{a+\varepsilon} and W=U baε W = U_{b - a - \varepsilon} for sufficiently small ε\varepsilon. Bridges’ counterexample showing that metric locatedness is not a uniform invariant means that this notion of uniform locatedness is strictly stronger than metric locatedness, since it implies metric locatedness with respect to all metrics inducing the same uniformity. Moreover, I think I can prove that if SS is uniformly located in this sense then so is N[S]N[S] for any entourage NN, since U[N[S]]=(UN)[S]U[N[S]] = (U\circ N)[S] and if WVUW\circ V\subseteq U then W(VN)(UN)W\circ (V\circ N) \subseteq (U\circ N).

    Unfortunately, this condition is much too strong. If the points of the discrete uniform space 22 are located in this sense, then LEM follows: let WW be the diagonal Δ 2\Delta_2, and U=V=Δ 2{(x,y)P}U = V = \Delta_2 \cup \{ (x,y) \mid P \} for any P:PropP:Prop. This suggests that we should only require it when the entourages U,V,WU,V,W are themselves located in some sense, but that’s getting circular. Hmm. Maybe WVUW\circ V \subseteq U should be replaced by ¬VU=X×X\neg V \cup U = X\times X, so that “point are located” becomes trivial? But then the proof of “SS located \Rightarrow N[S]N[S] located” fails, which is what I really wanted.

    Actually, I guess “SS located \Rightarrow N[S]N[S] located” should also only be expected when NN is located (if then), since if SS is a point and N=U{(x,y)P}N = U \cup \{ (x,y) \mid P \} then N[S]N[S] is probably not located.

    Have to go, so I’m posting this kind of half-baked, but maybe someone will have helpful thoughts…

    • CommentRowNumber6.
    • CommentAuthorTobyBartels
    • CommentTimeDec 11th 2016
    • (edited Dec 11th 2016)

    Maybe we just want a base of entourages that has this property, recognizing that general entourages (being completely arbitrary supersets of basic entourages) can be unlocatably weird.

    • CommentRowNumber7.
    • CommentAuthorMike Shulman
    • CommentTimeDec 12th 2016

    Yes, I thought of that too. By itself, it’s not as strong as what I wanted, e.g. if we use the base of entourages {(x,y)d(x,y)<1/2 n}\{(x,y) \mid d(x,y) \lt 1/2^n \} then we can’t conclude, for instance, that d(x,S)<1d(x,S)>34d(x,S)\lt 1 \vee d(x,S)\gt \frac{3}{4}. Even requiring the base to be closed under \circ doesn’t quite give this, but it might still give an interesting notion (and I think closure under \circ is necessary for the SN[S]S\rightsquigarrow N[S] step anyway).

    This is again weaker than metric locatedness, which at the moment I’m inclined to regard as a good thing: in Bridges’ example one might object that the set is only “accidentally” located with respect to one metric, but in fact I think I would argue that this set should be located (the metric space in question is itself weird, but the subset of it doesn’t involve any additional weirdness), and so the freedom to “choose an appropriate metric” is a good thing.

    It does still seem that “points are located” is an additional condition for this notion of locatedness, but at least it’s not an unreasonably strong condition. I hope it would be satisfied by all metrizable or even gaugeable uniform spaces, although closing the base up under \circ makes that not totally obvious.

    • CommentRowNumber8.
    • CommentAuthorMike Shulman
    • CommentTimeDec 12th 2016

    Hmm… maybe it is unreasonably strong. If qq ranges over a set of positive real numbers with infimum 00, then the closure under \circ of the entourage base U q={(x,y)d(x,y)<q}U_q = \{ (x,y) \mid d(x,y)\lt q\} for a metric space consists of the sets

    U q 0,,q n{(x,y)z 1,,z n,d(x,z 1)<q 0d(z 1,z 2)<q 1d(z n,y)<q n}U_{q_0,\dots,q_n} \{ (x,y) \mid \exists z_1,\dots,z_n, d(x,z_1)\lt q_0 \wedge d(z_1,z_2)\lt q_1 \wedge \cdots \wedge d(z_n,y) \lt q_n \}

    where U q n+1,,q mU q 0,,q n=U q 0,,q mU_{q_{n+1},\dots,q_m} \circ U_{q_0,\dots,q_n} = U_{q_0,\dots,q_m}. There are canonical inclusions U q 0,,q mU r 0,,r kU_{q_0,\dots,q_m} \subseteq U_{r_0,\dots,r_k} whenever we have an order-preserving map σ:{0,,m}{0,,k}\sigma:\{0,\dots,m\} \to \{0,\dots,k\} such that σ(j)=iq jr i\sum_{\sigma(j)=i} q_j \le r_i for all ii. But even if an inclusion U q n+1,,q mU q 0,,q nU r 0,,r kU_{q_{n+1},\dots,q_m} \circ U_{q_0,\dots,q_n} \subseteq U_{r_0,\dots,r_k} is of this form, I don’t see how to show (x,y)U r 0,,r k(x,y)\in U_{r_0,\dots,r_k} or (x,y)U q 0,,q n(x,y)\notin U_{q_0,\dots,q_n}. Even in the simple case U rU qU q,rU_r \circ U_q \subseteq U_{q,r}, we would be claiming that for any x,yx,y either there is a zz with d(x,z)<qd(x,z)\lt q and d(z,y)<rd(z,y)\lt r, or d(x,y)qd(x,y)\ge q. We can say that either d(x,y)<q+rd(x,y)\lt q+r or d(x,y)qd(x,y)\ge q, but in a general metric space there’s no reason the former should be “subdividable” into the existence of a zz.

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)