Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
28 changes: 0 additions & 28 deletions content/Meas_not_regular.md

This file was deleted.

41 changes: 41 additions & 0 deletions content/separated_objects_quasitopos.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,41 @@
---
title: Quasitopoi of Separated Objects
description: Some results concerning the full subcategory of separated objects for a Lawvere-Tierney topology on a topos
author: Daniel Schepler
---

## Results Concerning Quasitopoi of Separated Objects

Let $\T$ be an elementary topos with a <a href="https://ncatlab.org/nlab/show/Lawvere-Tierney+topology" target="_blank">Lawvere-Tierney topology</a> $j$. Recall that by definition, a subobject $V \hookrightarrow U$ is $j$-dense if $\cl_{j,U}(V) = U$, and it is $j$-closed if $\cl_{j,U}(V) = V$. Also, an object $X$ is $j$-separated if for each $j$-dense subobject $V \hookrightarrow U$, the composition function $\Hom(U, X) \to \Hom(V, X)$ is injective; or equivalently, if the diagonal subobject $\Delta_X : X \hookrightarrow X \times X$ is $j$-closed. In terms of the internal language of a topos, this last condition is equivalent to $1 \models \forall x, y : X, j(x=y) \rightarrow x=y$.

### Reflective Subcategory

::: Lemma 1
Let $\T$ be an elementary topos with a <a href="https://ncatlab.org/nlab/show/Lawvere-Tierney+topology" target="_blank">Lawvere-Tierney topology</a> $j$. Then for each object $X$ of $\T$, $\cl_{j,X\times X}(\Delta_X)$ is a congruence whose quotient $X_{\sep}$ is $j$-separated. Moreover, this operation forms a reflector functor making the full subcategory $\Sep(j)$ of $j$-separated objects a reflective category; and this reflector functor preserves finite products and monomorphisms.
:::

_Proof._ For generalized elements $x, y \in X(U)$, we see that $(x, y) \in \cl_{j,X\times X}(\Delta_X)$ if and only if $U \models j(x=y)$. This makes it easy to show the closure of the diagonal is a congruence: for reflexivity, we certainly have $U \models x=x$ and $(x=x) \le j(x=x)$. For symmetry, if $U \models j(x=y)$, then $(x=y) = (y=x)$, so $j(x=y) = j(y=x)$ and $U \models j(y=x)$ also. For transitivity, suppose $U \models j(x=y)$ and $U \models j(y=z)$. Then $U \models j(x=y) \land j(y=z)$, whereas $j(x=y) \land j(y=z) = j(x=y \land y=z)$. Also, $(x=y \land y=z) \le (x=z)$, so $j(x=y \land y=z) \le j(x=z)$, implying that $U \models j(x=z)$.

To see the quotient is separated, suppose we have generalized elements $\bar x, \bar y \in X_{sep}(U)$ such that $U \models j(\bar x = \bar y)$. Then since the quotient morphism $p : X \to X_{sep}$ is an epimorphism, there exists an epimorphism $\alpha : U' \to U$ and generalized elements $x, y \in X(U')$ such that $p(x) = \alpha^* \bar x$ and $p(y) = \alpha^* \bar y$. Since the quotient is effective, we have $\alpha^*(\bar x = \bar y) = (p(x) = p(y)) = j(x=y)$. Therefore, $\top = \alpha^*(j(\bar x = \bar y)) = j(\alpha^*(\bar x = \bar y)) = j(j(x=y)) = j(x=y)$, implying that $\alpha^*(\bar x = \bar y) = \top$. Since $\alpha$ is an epimorphism, that implies $\bar x = \bar y$. We have thus established that $1 \models \forall \bar x, \bar y : X_{\sep}, j(\bar x=\bar y) \rightarrow \bar x = \bar y$.

To see this forms a left adjoint to the forgetful functor $\Sep(j) \to \T$, suppose we have a morphism $f : X \to Y$ in $\T$ with $Y$ $j$-separated. Then $f$ factors through the projection morphism $p_X : X \to X_{\sep}$ if and only if $1 \models \forall x_1, x_2 : X, j(x_1 = x_2) \rightarrow f(x_1) = f(x_2)$. However, for generalized elements $x_1, x_2 \in X(U)$, we certainly have $(x_1 = x_2) \le (f(x_1) = f(x_2)$, so $j(x_1 = x_2) \le j(f(x_1) = f(x_2)) \le (f(x_1) = f(x_2))$ by the assumption that $Y$ is $j$-separated. We thus see $f$ does factor through the projection morphism $p_X : X \to X_{\sep}$, and the uniqueness of the factor $X_{\sep} \to Y$ is automatic since $p_X$ is an epimorphism. This implies that there is a unique functor extension of $X \mapsto X_{\sep}$ which forms a left adjoint, with the projection morphisms $p_X : X \to X_{\sep}$ serving as the unit of the adjunction.

To see that the reflector preserves monomorphisms, suppose we have a monomorphism $f : X \to Y$. Now, suppose we have two generalized elements $\bar x_1, \bar x_2 \in X_{\sep}(U)$ such that $f_{\sep}(\bar x_1) = f_{\sep}(\bar x_2)$ in $Y_{\sep}(U)$. Since $X \to X_{\sep}$ is an epimorphism, there exists an epimorphism $\alpha : U' \to U$ and $x_1, x_2 \in X(U')$ such that $\alpha^* \bar x_i = p_X(x_i)$, $i=1,2$. Then since $f_{\sep}(\alpha^* \bar x_1) = f_{\sep}(\alpha^* \bar x_2)$, we have $p_Y(f(x_1)) = p_Y(f(x_2))$. It follows that $U' \models j(f(x_1) = f(x_2))$. However, since $f$ is a monomorphism, $(f(x_1) = f(x_2)) \le (x_1 = x_2)$, so $j(f(x_1) = f(x_2)) \le j(x_1 = x_2)$, so $U' \models j(x_1 = x_2)$ also. Therefore, $p_X(x_1) = p_X(x_2)$, so $\alpha^* \bar x_1 = \alpha^* \bar x_2$. Since $\alpha$ is an epimorphism, this implies $\bar x_1 = \bar x_2$.

Finally, to see that the reflector preserves finite products, suppose we have objects $X_1, \ldots, X_n$ of $\T$. Then we have an epimorphism $\prod_{i=1}^n X_i \to \prod_{i=1}^n (X_i)_{\sep}$. In order to show the reflector preserves this finite product, it suffices to show the kernel pair of this epimorphism agrees with the $j$-closure of the diagonal on $\prod_{i=1}^n X_i$. That kernel pair is precisely the relation of pairs $(x = (x_1, \ldots, x_n), x' = (x_1', \ldots, x_n'))$ such that $\bigwedge_{i=1}^n j(x_i = x_i')$. However, since $j(\top) = \top$ and $j(p \land q) = j(p) \land j(q)$, we see $j$ preserves finite conjunctions, so this is equivalent to $j \left( \bigwedge_{i=1}^n (x_i = x_i') \right) = j(x = x')$, as required. <span class="qed">$\square$</span>

### Special Morphisms

::: Lemma 2
Let $\T$ be an elementary topos with a <a href="https://ncatlab.org/nlab/show/Lawvere-Tierney+topology" target="_blank">Lawvere-Tierney topology</a> $j$. Then in the full subcategory $\Sep(j)$ of $j$-separated objects:<br>
(a) The monomorphisms are the morphisms whose image in $\T$ are monomorphisms.<br>
(b) The epimorphisms are the morphisms whose image in $\T$ are $j$-dominant (i.e. the image calculated in $\T$ is a $j$-dense subobject of the codomain).<br>
(c) The regular monomorphisms are the morphisms whose image in $\T$ are $j$-closed monomorphisms.<br>
(d) The regular epimorphisms are the morphisms whose image in $\T$ are epimorphisms.
:::

_Proof._ By the above, $\Sep(j)$ is a reflective subcategory of $\T$, where the reflector takes an object $X$ to the quotient $X_{\sep}$ of $X$ by the congruence defined by the $j$-closure of the diagonal in $X\times X$. Also recall that the equalizer of $j, \id : \Omega_\T \rightrightarrows \Omega_\T$ is a $j$-separated object $\Omega_j$ which serves as the regular subobject classifier in $\Sep(j)$, and since $j$ is idempotent, this can also be described as the image (in $\T$) of $j$.<br>
(a) ($\Rightarrow$) This follows from the fact that $\Sep(j)$ is a reflective subcategory of $\T$. ($\Leftarrow$) This is trivial for any subcategory.<br>
(b) ($\Rightarrow$) Given a morphism $f : X \to Y$, form the image $\im(f)$ in $\T$, which corresponds to a morphism $\chi_{\im(f)} : Y \to \Omega_\T$. Then $j \circ \chi_{\im(f)} \circ f = j\circ \top_X = \top_X = \top_Y \circ f$ as morphisms $Y \to \Omega_j$, so if $f$ is an epimorphism in $\Sep(j)$, then we conclude $j \circ \chi_{\im(f)} = \top_Y$. However, $j \circ \chi_{\im(f)} : Y \to \Omega_j \hookrightarrow \Omega$ is the characteristic morphism of the $j$-closure of $\im(f)$, so we conclude that $j$-closure is all of $Y$. ($\Leftarrow$) Given a morphism $f : X \to Y$ of $j$-separated objects whose image in $\T$ is $j$-dense, suppose we have two morphisms $g, h : Y \rightrightarrows Z$ with $g \circ f = h \circ f$. Then since $Z$ is $j$-separated, the equalizer of $g$ and $h$ is $j$-closed; it also contains the image of $f$ and thus is $j$-dense. We conclude that the equalizer is all of $Y$.<br>
(c) ($\Rightarrow$) Any equalizer in $\T$ of $f, g : X \rightrightarrows Y$ with $Y$ $j$-separated is a $j$-closed subobject of $X$. If $X$ is $j$-separated as well, then that equalizer subobject is automatically separated, and agrees with the equalizer in $\Sep(j)$. ($\Leftarrow$) For a $j$-closed subobject $f : X \hookrightarrow Y$, we see that the characteristic morphism in $\T$, $\chi_X : Y \to \Omega_\T$, factors through $\Omega_j$. Now $X$ is the equalizer of $\chi_X, \top : Y \rightrightarrows \Omega_j$.<br>
(d) ($\Rightarrow$) We can calculate the coequalizer of $f, g : X \rightrightarrows Y$ in $\Sep(j)$ by taking the coequalizer $Z$ in $\T$ and then applying the reflector to get $Z_{sep}$. We see that both $Y \to Z$ and $Z \to Z_{sep}$ are epimorphisms in $\T$. ($\Leftarrow$) Suppose $f : X \to Y$ is an epimorphism in $\T$ of $j$-separated objects. Then the subcategory inclusion functor preserves the kernel pair $X \times_Y X \rightrightarrows X$, and since $f$ is a regular epimorphism in $\T$, this kernel pair has coequalizer $f : X \to Y$ in $\T$. Since $Y$ was already $j$-separated, the kernel pair also has coequalizer $f : X \to Y$ in $\Sep(j)$. <span class="qed">$\square$</span>
3 changes: 0 additions & 3 deletions databases/catdat/data/categories/Cat.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -43,9 +43,6 @@ unsatisfied_properties:
- property: cogenerating set
proof: 'Assume that $S$ is a cogenerating set in $\Cat$. Then one checks that the set of monoids $\{\End(X) : X \in \C \in S\}$ is a cogenerating set in <a href="/category/Mon">$\Mon$</a>, which we know does not exist.'

- property: regular
proof: See Example 3.14 at the <a href="https://ncatlab.org/nlab/show/regular+category" target="_blank">nLab</a>.

- property: coregular
proof: 'We already know that <a href="/category/Mon">$\Mon$</a> is not coregular, in fact there is a regular monomorphism $M \to N$ of monoids and a morphism $M \to K$ such that $K \to K \sqcup_M N$ is not a monomorphism. The delooping functor $B : \Mon \to \Cat$ has a left adjoint (<a href="https://math.stackexchange.com/questions/574745" target="_blank">MSE/574745</a>), hence it preserves regular monomorphisms. It also preserves pushouts (<a href="https://math.stackexchange.com/questions/5130854" target="_blank">MSE/5130854</a>), and it reflects monomorphisms since it is faithful. Therefore, $B(M) \to B(N)$ provides the desired counterexample of a non-stable regular monomorphism of categories.'

Expand Down
1 change: 1 addition & 0 deletions databases/catdat/data/categories/LRS_R.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -62,6 +62,7 @@ unsatisfied_properties:

- property: regular
proof: This is Corollary 4(c) <a href="/content/Top-embeds-in-LRS">here</a>.
check_redundancy: false

- property: cofiltered-limit-stable epimorphisms
proof: This is Corollary 4(d) <a href="/content/Top-embeds-in-LRS">here</a>.
Expand Down
6 changes: 0 additions & 6 deletions databases/catdat/data/categories/Meas.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -56,9 +56,6 @@ unsatisfied_properties:
- property: balanced
proof: Take a set $X$ with two different $\sigma$-algebras $\A \subset \B$ (for example, $\A = \{\varnothing,X\}$ and $\B = P(X)$ when $X$ has at least $2$ elements), then the identity map $(X,\B) \to (X,\A)$ provides a counterexample.

- property: Malcev
proof: Use that <a href="/category/Set">$\Set$</a> is not Malcev and endow sets with the trivial $\sigma$-algebra.

- property: cartesian filtered colimits
proof: See <a href="https://math.stackexchange.com/questions/5027218" target="_blank">MSE/5027218</a>.

Expand All @@ -68,9 +65,6 @@ unsatisfied_properties:
- property: effective cocongruences
proof: 'The proof is similar to the one for <a href="/category/Top">$\Top$</a>: Use the trivial $\sigma$-algebra on a two-point set.'

- property: regular
proof: A proof can be found <a href="/content/Meas_not_regular">here</a>.

special_objects:
initial object:
description: empty set with the unique $\sigma$-algebra
Expand Down
3 changes: 0 additions & 3 deletions databases/catdat/data/categories/Met_oo.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -60,9 +60,6 @@ unsatisfied_properties:
- property: effective cocongruences
proof: The same counterexample as for <a href="/category/Met">$\Met$</a> works here. The difference in this case is that a binary copower of two copies of $(0,1)$ does exist in $\Met_\infty$. However, this would assign a distance of $\infty$ between points in $(-1,0)$ and points in $(0,1)$, which does not agree with the chosen subspace metric on $(-1,0) \cup (0,1)$.

- property: regular
proof: We can take the same counterexample as for <a href="/category/PMet">$\PMet$</a>.

special_objects:
initial object:
description: empty metric space
Expand Down
3 changes: 0 additions & 3 deletions databases/catdat/data/categories/Pos.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -45,9 +45,6 @@ unsatisfied_properties:
- property: balanced
proof: The inclusion $\{0,1\} \to \{0 < 1\}$ provides a counterexample (where in the domain there is no relation between $0$ and $1$).

- property: regular
proof: See Example 3.14 at the <a href="https://ncatlab.org/nlab/show/regular+category" target="_blank">nLab</a>.

- property: Malcev
proof: 'Consider the subposet $\{(a,b) : a \leq b \}$ of $\IN^2$.'

Expand Down
6 changes: 0 additions & 6 deletions databases/catdat/data/categories/Prost.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -41,18 +41,12 @@ satisfied_properties:
proof: The set $\{0,1\}$ with the chaotic preorder $(0 \leq 1$, $1 \leq 0)$ is a regular subobject classifier since order-preserving maps $P \to \{0,1\}$ correspond to subsets of $P$.

unsatisfied_properties:
- property: regular
proof: See Example 3.14 at the <a href="https://ncatlab.org/nlab/show/regular+category" target="_blank">nLab</a>.

- property: balanced
proof: The inclusion $\{0,1\} \to \{0 < 1\}$ provides a counterexample (where in the domain there is no relation between $0$ and $1$).

- property: skeletal
proof: This is trivial.

- property: Malcev
proof: 'Consider the subproset $\{(a,b) : a \leq b \}$ of $\IN^2$.'

- property: co-Malcev
proof: 'See <a href="https://mathoverflow.net/questions/509552">MO/509552</a>: Consider the forgetful functor $U : \Prost \to \Set$ and the relation $R \subseteq U^2$ defined by $R(A) \coloneqq \{(a,b) \in U(A)^2 : a \leq b\}$. Both are representable: $U$ by the singleton preordered set and $R$ by $\{0 \leq 1 \}$. It is clear that $R$ is reflexive, but not symmetric.'

Expand Down
3 changes: 0 additions & 3 deletions databases/catdat/data/categories/Rng.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -55,9 +55,6 @@ unsatisfied_properties:
- property: CSP
proof: Assume that $\coprod_n \IZ \to \prod_n \IZ$ is an epimorphism in $\Rng$. Then $((\coprod_n \IZ)^+)^{\ab} \to \prod_n \IZ$ would be an epimorphism in $\CRing$, where $(-)^+$ denotes the unitalization and $(-)^{\ab}$ the abelianization. But if $R \to S$ is an epimorphism of commutative rings, then $\card(S) \leq \card(R)$ by <a href="https://stacks.math.columbia.edu/tag/04W0" target="_blank">SP/04W0</a>. Since $((\coprod_n \IZ)^+)^{\ab}$ is countable and $\prod_n \IZ$ is not, we get a contradiction.

- property: regular subobject classifier
proof: 'Assume that $\Rng$ has a subobject classifier $\Omega$. Since $0$ is a zero object, every regular subobject $R \subseteq S$ would be the kernel of some homomorphism $S \to \Omega$. In particular, it would be an ideal. Now take any pair of homomorphisms $f,g : S \rightrightarrows T$ in $\Ring$. Their equalizer $R \subseteq S$ is also the equalizer in $\Rng$, and it contains $1 \in S$. If it was an ideal, then $R = S$, and hence $f = g$, which is absurd.'

- property: coregular
proof: 'We can copy the proof for <a href="/category/Ring">$\Ring$</a>. In short, the inclusion of diagonal matrices $\IQ^2 \hookrightarrow M_2(\IQ)$ is a regular monomorphism, but becomes zero after taking the pushout with $p_1 : \IQ^2 \twoheadrightarrow \IQ$ because $M_2(\IQ)$ is simple.'

Expand Down
Loading