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
12 changes: 9 additions & 3 deletions content/generator_construction.md
Original file line number Diff line number Diff line change
@@ -1,13 +1,19 @@
---
title: Construction of Generators
description: How to construct a generator from a generating set
author: Martin Brandenburg
authors:
- Martin Brandenburg
- Daniel Schepler
---

## Construction of Generators

::: Lemma
In a category let $S$ be a generating set which is [strongly connected](/category-property/strongly_connected) (between any two objects in $S$ there is a morphism). If the coproduct $U \coloneqq \coprod_{G \in S} G$ exists, then it is a generator.
In a category let $S$ be a generating set which is [strongly connected](/category-property/strongly_connected) (between any two objects in $S$ there is a morphism). If the coproduct $U \coloneqq \coprod_{G \in S} G$ exists, then it is a generator. Moreover, if $S$ is an extremal generating set, then $U$ is an extremal generator.
:::

_Proof._ This is a straight forward generalization of [this result](/category-implication/generator_via_coproduct). We remark that the assumption about $S$ implies that each inclusion $G \to U$ has a left inverse. Now let $f,g : A \rightrightarrows B$ be two morphisms with $f h = g h$ for all $h : U \to A$. If $G \in S$, any morphism $G \to A$ extends to $U$ by our preliminary remark. Thus, $fh = gh$ holds for all $h : G \to A$ and $G \in S$. Since $S$ is a generating set, this implies $f = g$. <span class="qed">$\square$</span>
_Proof._ We remark that the assumption on $S$ implies that each coprojection $i_G : G \to U$ has a left inverse. Now let $f,g : A \rightrightarrows B$ be two morphisms with $f \circ \bar a = g \circ \bar a$ for all $\bar a : U \to A$. If $G \in S$, any morphism $G \to A$ extends to $U$ by our preliminary remark. Thus, $f \circ a = g \circ a$ holds for all morphisms $a : G \to A$ with $G \in S$. Since $S$ is a generating set, this implies $f = g$.

Similarly, for the case where $S$ is an extremal generating set, suppose we have a morphism $f : A \to B$ such that $f \circ {-} : \Hom(U, A) \to \Hom(U, B)$ is a bijection. In particular, because it is injective and $U$ is a generator, we can conclude that $f$ is a monomorphism, so $f \circ {-} : \Hom(G, A) \to \Hom(G, B)$ is injective for each $G \in S$. Now suppose $b \in \Hom(G, B)$ for $G \in S$. Then $b$ extends to a morphism $\bar b : U \to B$. By assumption, there exists $\bar a : U \to A$ such that $f \circ \bar a = \bar b$. Composing with the coprojection $i_G : G \to U$, we see
$$f \circ \bar a \circ i_G = \bar b \circ i_G = b.$$
This shows that $f \circ {-} : \Hom(G, A) \to \Hom(G, B)$ is also surjective for each $G \in S$. Since $S$ is an extremal generating set, this implies $f$ is an isomorphism. <span class="qed">$\square$</span>
21 changes: 21 additions & 0 deletions content/thin_extremal_generator.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
---
title: Thin Category with an Extremal Generator
description: A result restricting which thin categories can have an extremal generator
author: Daniel Schepler
---

# Thin Category with an Extremal Generator

::: Lemma
Suppose $G$ is an extremal generator of a thin category. Then for any object $X$, either $X \cong G$ or every morphism with codomain $X$ is an isomorphism.
:::

_Proof._ Since the category is thin, $\Hom(G, X)$ is either a singleton or empty. In the first case, let $f \in \Hom(G, X)$. Then $f \circ {-} : \Hom(G, G) \to \Hom(G, X)$ is automatically a bijection since $\Hom(G, G) = \{ \id_G \}$ is also a singleton, implying that $f$ is an isomorphism.

In the second case, suppose we have a morphism $g : Y \to X$. Then $g \circ {-} : \Hom(G, Y) \to \Hom(G, X)$ is a function with empty codomain, so it is automatically a bijection, implying that $g$ is an isomorphism. <span class="qed">$\square$</span>

::: Corollary
A poset whose corresponding thin category has an extremal generator can contain at most one non-minimal element.
:::

_Proof._ In a thin category coming from a poset, the second condition in the previous lemma is equivalent to the corresponding element of the poset being minimal. <span class="qed">$\square$</span>
4 changes: 2 additions & 2 deletions database/data/categories/Ab_fg.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -21,8 +21,8 @@ satisfied_properties:
- property: abelian
proof: This follows from the fact for abelian groups and the fact that subgroups of finitely generated abelian groups are also finitely generated.

- property: generator
proof: The group $\IZ$ is a generator since it represents the forgetful functor to $\Set$.
- property: extremal generator
proof: The group $\IZ$ is an extremal generator since it represents the forgetful functor to $\Set$ which is faithful and conservative.

- property: essentially countable
proof: Every finitely generated abelian group is isomorphic to a group of the form $\IZ^n / U$, where $n \in \IN$ and $U$ is a subgroup of $\IZ^n$. Since $\IZ^n$ is Noetherian as a $\IZ$-module, $U$ is finitely generated, hence the category $\Ab_\fg$ has only countably many objects up to isomorphism. Furthermore, for any objects $A \cong \IZ^n / U$ and $B \cong \IZ^m / T$, the hom-set $\Hom(A,B)$ is countable. Indeed, precomposition with the quotient map yields an injection $\Hom(A,B) \hookrightarrow \Hom(\IZ^n, B) \cong B^n$, and $B^n$ is countable.
Expand Down
5 changes: 3 additions & 2 deletions database/data/categories/Ban.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -20,8 +20,9 @@ satisfied_properties:
proof: The trivial Banach space $\{0\}$ is a zero object.
check_redundancy: false

- property: cogenerator
proof: The Hahn-Banach theorem implies that $\IC$ is a cogenerator.
- property: extremal cogenerator
proof: >-
The Hahn-Banach theorem implies that $\IC$ is a cogenerator. We claim that it is in fact an extremal cogenerator. Thus, suppose $f : X \to Y$ is a morphism such that ${-} \circ f : \Hom(Y, \IC) \to \Hom(X, \IC)$ is bijective on the underlying sets. Then for any $x \in X$, by the Hahn-Banach theorem, there exists $\varphi \in X^*$ such that $|\varphi| = 1$ and $\varphi(x) = |x|$. Since $|\varphi| = 1$, we see that $\varphi$ is a morphism $X \to \IC$ in $\Ban$; so by the assumption, there exists a morphism $\psi : Y \to \IC$ such that $\varphi = \psi \circ f$. Therefore, $|x| = \psi(f(x)) \le |f(x)|$; and conversely, since $f$ is a morphism, $|f(x)| \le |x|$. This shows that $f$ is isometric and therefore a regular monomorphism (see below). On the other hand, since $\IC$ is a cogenerator and ${-} \circ f$ is injective, we have $f$ is also an epimorphism. Hence, $f$ is an isomorphism.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Then for any $x \in X$, by the Hahn-Banach theorem, there exists $\varphi \in X^*$ such that $|\varphi| = 1$ and $\varphi(x) = |x|$.

As commented before, this is not true for, say, $X = 0$. I think we should assume $x \neq 0$.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

As commented before, the norms are missing in

Therefore, $|x| = \psi(f(x)) \le |f(x)|$;


- property: CIP
proof: This is immediate from the concrete description of coproducts and products.
Expand Down
7 changes: 5 additions & 2 deletions database/data/categories/Cat.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -27,8 +27,11 @@ satisfied_properties:
- property: semi-strongly connected
proof: Every non-empty category is weakly terminal (by using constant functors).

- property: generator
proof: 'The <a href="/category/walking_morphism">walking morphism</a> $I$ is a generator: Assume that $F,G : \C \rightrightarrows \D$ are functors that agree when being precomposed with any functor from $I$. This means that $F(f) = G(f)$ for all morphisms $f : X \to Y$ in $\C$. By comparing the domains and applying this to $f = \id_X$, we see that $F(X) = G(X)$ for all objects $X$. And we just saw that $F,G$ also agree on morphisms.'
- property: extremal generator
proof: >-
The <a href="/category/walking_morphism">walking morphism</a> $I$ is a generator: Assume that $F,G : \C \rightrightarrows \D$ are functors that agree when being precomposed with any functor from $I$. This means that $F(f) = G(f)$ for all morphisms $f : X \to Y$ in $\C$. By comparing the domains and applying this to $f = \id_X$, we see that $F(X) = G(X)$ for all objects $X$. And we just saw that $F,G$ also agree on morphisms.

In fact, $I$ is an extremal generator: suppose $F : \C \to \D$ is a functor which induces a bijection $\Mor(\C) \to \Mor(\D)$. By considering the images of identity morphisms in $\C$, we see that $F$ is injective on objects; and then by considering the preimages of identity morphisms in $\D$, we see that $F$ is surjective on objects. By assumption, $F$ is also bijective on morphisms, so $F$ is an isomorphism of categories.

- property: infinitary extensive
proof: '[Sketch] This is straight forward from the fact that <a href="/category/Set">$\Set$</a> is infinitary extensive: A functor $\C \to \coprod_i \D_i$ yields full subcategories $\C_i \subseteq \C$ (the preimages of $\D_i)$ with $\C = \coprod_i \C_i$.'
Expand Down
8 changes: 4 additions & 4 deletions database/data/categories/Delta.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -33,11 +33,11 @@ satisfied_properties:
- property: strongly connected
proof: For all $n,m$ there are morphisms $[n] \to [0] \to [m]$.

- property: generator
proof: The ordered set $[0] = \{0\}$ is a generator.
- property: extremal generator
proof: The ordered set $[1] = \{0 < 1\}$ is an extremal generator, even for <a href="/category/PreOrd">$\PreOrd$</a>.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I assume we are using the following fact?

If C is a full subcategory of D and X is an object in C which is an extremal generator in D, then it is an extremal generator in C.

This is trivial, but it is used so much (see the other comments below) that maybe we can add it to subcategories.md. Either way, I would like to mention this somewhere.


- property: cogenerator
proof: The ordered set $[1] = \{0 < 1\}$ is a cogenerator, even for <a href="/category/Pos">$\Pos$</a>.
- property: extremal cogenerator
proof: The ordered set $[1] = \{0 < 1\}$ is an extremal cogenerator, even for <a href="/category/Pos">$\Pos$</a>.

- property: skeletal
proof: 'If $f : [n] \to [m]$ is an isomorphism, then $n + 1 = m + 1$ by comparing the cardinalities, hence $n = m$.'
Expand Down
4 changes: 2 additions & 2 deletions database/data/categories/FI.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -26,8 +26,8 @@ satisfied_properties:
- property: left cancellative
proof: This is trivial.

- property: generator
proof: The one-point set is a generator since it represents the forgetful functor $\FI \to \Set$.
- property: extremal generator
proof: The one-point set is a generator since it represents the forgetful functor $\FI \to \Set$, which is faithful and conservative.

- property: essentially countable
proof: Every finite set is isomorphic to some $\{1,\dotsc,n\}$ for some $n \in \IN$.
Expand Down
7 changes: 5 additions & 2 deletions database/data/categories/FS.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -28,8 +28,11 @@ satisfied_properties:
- property: right cancellative
proof: This is trivial.

- property: cogenerator
proof: 'We prove that $\{0,1\}$ is a cogenerator: The surjective maps $X \to \{0,1\}$ correspond to the non-empty proper subsets of $X$. If $a,b \in X$ are elements that have the same image under each surjective map $X \to \{0,1\}$, it therefore means that they lie in the same non-empty proper subsets of $X$. This implies $a=b$: If $X = \{a\}$, this is trivial. Otherwise, use the subset $\{a\}$.'
- property: extremal cogenerator
proof: >-
We prove that $\{0,1\}$ is an extremal cogenerator. First, to prove it is a cogenerator: The surjective maps $X \to \{0,1\}$ correspond to the non-empty proper subsets of $X$. If $a,b \in X$ are elements that have the same image under each surjective map $X \to \{0,1\}$, it therefore means that they lie in the same non-empty proper subsets of $X$. This implies $a=b$: If $X = \{a\}$, this is trivial. Otherwise, use the subset $\{a\}$.

Now, suppose we have a surjective morphism $f : X \to Y$ of finite sets such that ${-} \circ f : \Hom(Y, \{0,1\}) \to \Hom(X, \{0,1\})$ is bijective. That means that $f^* : P(Y) \to P(X)$ is bijective on non-empty subsets, and it certainly also maps $\varnothing \mapsto \varnothing$. Therefore, since the <a href="/functor/power_set_contravariant">contravariant powerset functor</a> is conservative, that implies $f$ is an isomorphism.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
Now, suppose we have a surjective morphism $f : X \to Y$ of finite sets such that ${-} \circ f : \Hom(Y, \{0,1\}) \to \Hom(X, \{0,1\})$ is bijective. That means that $f^* : P(Y) \to P(X)$ is bijective on non-empty subsets, and it certainly also maps $\varnothing \mapsto \varnothing$. Therefore, since the <a href="/functor/power_set_contravariant">contravariant powerset functor</a> is conservative, that implies $f$ is an isomorphism.
Now, suppose we have a surjective map $f : X \to Y$ of finite sets such that ${-} \circ f : \Hom(Y, \{0,1\}) \to \Hom(X, \{0,1\})$ is bijective. That means that $f^* : P(Y) \to P(X)$ is bijective on non-empty proper subsets, and it certainly also maps $\varnothing \mapsto \varnothing$ and $Y \mapsto X$. Therefore, since the <a href="/functor/power_set_contravariant">contravariant powerset functor</a> is conservative, that implies $f$ is an isomorphism.


- property: coequalizers
proof: We construct coequalizers as in <a href="/category/FinSet">$\FinSet$</a> (or $\Set$) and observe that the universal property still holds when we restrict to surjective maps.
Expand Down
4 changes: 4 additions & 0 deletions database/data/categories/FiltVect.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -47,6 +47,10 @@ satisfied_properties:

- property: cogenerator
proof: It is straightforward to check that the vector space $K$ equipped with the maximal filtration $F^n(K) \coloneqq K$ is a cogenerator.
check_redundancy: false

- property: extremal cogenerating set
proof: 'Let $K_n$ denote the vector space $K$ equipped with the filtration such that $F_m(K) = K$ for $m < n$ and $F_m(K) = 0$ for $m \ge n$. Then $K_n \hookrightarrow (K, n \mapsto K)$ represents the functor $(V, F) \mapsto F_n(V)^{\perp} \subseteq V^*$. It is straightforward to check that the dual vector space functor along with these functors is a jointly conservative collection. Therefore, $\{ K_n : n \in \IZ \} \cup \{ (K, n \mapsto K) \}$ is an extremal cogenerating set.'

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I find this proof hard to follow. For example,

$K_n \hookrightarrow (K, n \mapsto K)$

is a monomorphism in FiltVect, so how can it represent a functor? I had to read this a couple of times to see what you mean.

Also, the notation $(K, n \mapsto K)$ is a bit inelegant. I think we can use $K_\infty$. This is consistent with the $K_n$.

Next, the notation $F_n(V)^{\perp}$ was not immediately clear to me. I think it is better to introduce it in a (half)sentence.

Finally, you claim that "the dual vector space functor along with these functors is a jointly conservative collection", but there is no proof. I think we should add it.

I have no doubts that everything is correct here, but can you please revise this proof and write it down more "slowly"?

PS: Please avoid force pushes and rewrites during the review process, since this makes it impossible for me to see what has changed. Of course it's OK when the review has finished and the PR is about to get merged.


- property: finitely accessible
proof: >-
Expand Down
8 changes: 4 additions & 4 deletions database/data/categories/FinOrd.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -34,11 +34,11 @@ satisfied_properties:
- property: semi-strongly connected
proof: Every non-empty totally ordered set is weakly terminal (by using constant maps).

- property: generator
proof: The one-point finite ordered set is a generator since it represents the forgetful functor $\FinOrd \to \Set$.
- property: extremal generator
proof: The same proof as for <a href="/category/PreOrd">$\PreOrd$</a> shows that $\{ 0<1 \}$ is an extremal generator of $\FinOrd$.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cannot we again say something like "even for ..."? Just like the proof below.


- property: cogenerator
proof: The ordered set $\{0 < 1\}$ is a cogenerator, even for <a href="/category/Pos">$\Pos$</a>.
- property: extremal cogenerator
proof: The ordered set $\{0 < 1\}$ is an extremal cogenerator, even for <a href="/category/Pos">$\Pos$</a>.

- property: equalizers
proof: Take the equalizer in $\FinSet$ and restrict the order.
Expand Down
8 changes: 4 additions & 4 deletions database/data/categories/FinSet.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -26,11 +26,11 @@ satisfied_properties:
- property: essentially countable
proof: Every finite set is isomorphic to some $\{1,\dotsc,n\}$ for some $n \in \IN$.

- property: generator
proof: The one-point set is a generator since it represents the forgetful functor $\FinSet \to \Set$.
- property: extremal generator
proof: The one-point set is an extremal generator since it represents the forgetful functor $\FinSet \to \Set$ which is faithful and conservative (in fact, fully faithful).

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cannot we just say that the one-point set is even an extremal generator in Set?

Let's add this info to Set.yaml, either as a comment or as a redundant property assignment. (Maybe I should add generators and extremal generators to special_objects.)


- property: cogenerator
proof: The two-element set is a cogenerator.
- property: extremal cogenerator
proof: The two-element set is an extremal cogenerator since it represents the contravariant power set functor $\FinSet^{\op} \to \Set$ which is faithful and conservative (see <a href="/functor/power_set_contravariant">contravariant power set functor</a>).

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cannot we just say that the two-element set is even an extremal cogenerator in Set?

Let's add this info to Set.yaml, either as a comment or as a redundant property assignment. (Maybe I should add generators and extremal generators to special_objects.)


- property: semi-strongly connected
proof: Every non-empty finite set is weakly terminal (by using constant maps).
Expand Down
4 changes: 2 additions & 2 deletions database/data/categories/FinVect_c.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -23,8 +23,8 @@ satisfied_properties:
- property: essentially countable
proof: Every object is isomorphic to $K^n$ for some $n \in \IN$, and $\Hom(K^n,K^m) \cong M_{m \times n}(K)$ is a countable set.

- property: generator
proof: The forgetful functor $\FinVect_K \to \Set$ is faithful and represented by $K$. Hence, $K$ is a generator.
- property: extremal generator
proof: The forgetful functor $\FinVect_K \to \Set$ is faithful and conservative, and it is represented by $K$. Hence, $K$ is a generator.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

First:

Suggested change
proof: The forgetful functor $\FinVect_K \to \Set$ is faithful and conservative, and it is represented by $K$. Hence, $K$ is a generator.
proof: The forgetful functor $\FinVect_K \to \Set$ is faithful and conservative, and it is represented by $K$. Hence, $K$ is an extremal generator.

But as before cannot we just say that $K$ is even an extremal generator in Vect and add this info to Vect.yaml?

Same remarks apply to the other entries of FinVect. Sorry for the duplication.


- property: split abelian
proof: This follows directly from the corresponding fact for <a href="/category/Vect">$\Vect_K$</a>.
Expand Down
4 changes: 2 additions & 2 deletions database/data/categories/FinVect_f.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -26,8 +26,8 @@ satisfied_properties:
- property: locally finite
proof: Each hom-set $\Hom(K^n,K^m) \cong M_{m \times n}(K)$ is finite by assumption.

- property: generator
proof: The forgetful functor $\FinVect_K \to \Set$ is faithful and represented by $K$. Hence, $K$ is a generator.
- property: extremal generator
proof: The forgetful functor $\FinVect_K \to \Set$ is faithful and conservative, and it is represented by $K$. Hence, $K$ is an extremal generator.

- property: split abelian
proof: This follows directly from the corresponding fact for <a href="/category/Vect">$\Vect_K$</a>.
Expand Down
4 changes: 2 additions & 2 deletions database/data/categories/FinVect_u.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -23,8 +23,8 @@ satisfied_properties:
- property: essentially small
proof: Every object is isomorphic to $K^n$ for some $n \in \IN$, and $\Hom(K^n,K^m) \cong M_{m \times n}(K)$ is a set.

- property: generator
proof: The forgetful functor $\FinVect_K \to \Set$ is faithful and represented by $K$. Hence, $K$ is a generator.
- property: extremal generator
proof: The forgetful functor $\FinVect_K \to \Set$ is faithful and conservative, and it is represented by $K$. Hence, $K$ is a generator.

- property: split abelian
proof: This follows directly from the corresponding fact for <a href="/category/Vect">$\Vect_K$</a>.
Expand Down
4 changes: 2 additions & 2 deletions database/data/categories/FreeAb.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -23,8 +23,8 @@ satisfied_properties:
- property: coproducts
proof: This is is because free abelian groups are closed under direct sums of abelian groups.

- property: generator
proof: As for <a href="/category/Ab">$\Ab$</a>, the group $\IZ$ is a generator.
- property: extremal generator
proof: As for <a href="/category/Ab">$\Ab$</a>, the group $\IZ$ is an extremal generator.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

As in my other comments: Let's try to make this clear that not an extra proof is needed, but that this can be deduced from the general lemma.


- property: cogenerator
proof: It is easy to check that $\IZ$ is a cogenerator for free abelian groups.
Expand Down
4 changes: 2 additions & 2 deletions database/data/categories/Grp_c.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -25,8 +25,8 @@ satisfied_properties:
proof: The trivial group is countable and is a zero object.
check_redundancy: false

- property: generator
proof: The countable group $\IZ$ is a generator because it represents the forgetful functor $\Grp_\c \to \Set$.
- property: extremal generator
proof: The countable group $\IZ$ is an extremal generator because it represents the forgetful functor $\Grp_\c \to \Set$ which is faithful and conservative.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

As before, let's rather use that Z is an extremal generator in Grp which happens to be in the subcategory.


- property: finite products
proof: This is because <a href="/category/Grp">$\Grp$</a> has finite (in fact, all) products, and $\Grp_\c \hookrightarrow \Grp$ is closed under finite products. This is because a finite product of countable sets is again countable.
Expand Down
Loading