Image (category theory)#Essential Image

In category theory, a branch of mathematics, the image of a morphism is a generalization of the image of a function.

General definition

Given a category C and a morphism f\colon X\to Y in C , the image{{Citation| last=Mitchell| first=Barry| title=Theory of categories|publisher=Academic Press| series=Pure and applied mathematics| isbn=978-0-12-499250-4| mr=0202787| year=1965| volume=17}} Section I.10 p.12

of f is a monomorphism m\colon I\to Y satisfying the following universal property:

  1. There exists a morphism e\colon X\to I such that f = m\, e.
  2. For any object I' with a morphism e'\colon X\to I' and a monomorphism m'\colon I'\to Y such that f = m'\, e', there exists a unique morphism v\colon I\to I' such that m = m'\, v.

Remarks:

  1. such a factorization does not necessarily exist.
  2. e is unique by definition of m monic.
  3. m'e'=f=me=m've, therefore e'=ve by m' monic.
  4. v is monic.
  5. m = m'\, v already implies that v is unique.

File:Image Theorie des catégories.png

The image of f is often denoted by \text{Im} f or \text{Im} (f).

Proposition: If C has all equalizers then the e in the factorization f= m\, e of (1) is an epimorphism.{{Citation| last=Mitchell| first=Barry| title=Theory of categories|publisher=Academic Press| series=Pure and applied mathematics| isbn=978-0-12-499250-4| mr=0202787| year=1965| volume=17}} Proposition 10.1 p.12

{{Math proof|

Let \alpha,\, \beta be such that \alpha\, e =\beta\, e, one needs to show that \alpha=\beta. Since the equalizer of (\alpha, \beta) exists, e factorizes as e= q\, e' with q monic. But then f= (m\, q)\, e' is a factorization of f with (m\, q) monomorphism. Hence by the universal property of the image there exists a unique arrow v: I \to Eq_{\alpha,\beta} such that m = m\,q\, v and since m is monic \text{id}_I = q\, v. Furthermore, one has m\, q = (m q v)\,q and by the monomorphism property of mq one obtains \text{id}_{Eq_{\alpha,\beta}}= v\, q.

File:E epimorphism.png

This means that I \equiv Eq_{\alpha,\beta} and thus that \text{id}_I = q\, v equalizes (\alpha, \beta), whence \alpha = \beta.

}}

Second definition

In a category C with all finite limits and colimits, the image is defined as the equalizer (Im,m) of the so-called cokernel pair (Y \sqcup_X Y, i_1, i_2), which is the cocartesian of a morphism with itself over its domain, which will result in a pair of morphisms i_1,i_2:Y\to Y\sqcup_X Y, on which the equalizer is taken, i.e. the first of the following diagrams is cocartesian, and the second equalizing.{{Citation|last= Kashiwara|first= Masaki| author1link = Masaki Kashiwara|author2link = Pierre Schapira (mathematician)|last2 = Schapira |first2= Pierre|title="Categories and Sheaves"|year=2006|publisher=Springer| series = Grundlehren der Mathematischen Wissenschaften| place= Berlin Heidelberg|pages = 113–114|volume=332}} Definition 5.1.1

File:Cokernel pair.png

File:Equalizer of the cokernel pair, diagram.png

Remarks:

  1. Finite bicompleteness of the category ensures that pushouts and equalizers exist.
  2. (Im,m) can be called regular image as m is a regular monomorphism, i.e. the equalizer of a pair of morphisms. (Recall also that an equalizer is automatically a monomorphism).
  3. In an abelian category, the cokernel pair property can be written i_1\, f = i_2\, f\ \Leftrightarrow\ (i_1 - i_2)\, f = 0 = 0\, f and the equalizer condition i_1\, m = i_2\, m\ \Leftrightarrow\ (i_1 - i_2)\, m = 0 \, m. Moreover, all monomorphisms are regular.

{{Math theorem|If f always factorizes through regular monomorphisms, then the two definitions coincide.

}}

{{Math proof|

First definition implies the second: Assume that (1) holds with m regular monomorphism.

  • Equalization: one needs to show that i_1\, m= i_2\, m . As the cokernel pair of f,\ i_1\, f= i_2\, f and by previous proposition, since C has all equalizers, the arrow e in the factorization f= m\, e is an epimorphism, hence i_1\, f= i_2\, f\ \Rightarrow \ i_1\, m= i_2\, m.
  • Universality: in a category with all colimits (or at least all pushouts) m itself admits a cokernel pair (Y \sqcup_{I}Y, c_1, c_2)

File:Cokernel pair m.png

:Moreover, as a regular monomorphism, (I,m) is the equalizer of a pair of morphisms b_1, b_2: Y \longrightarrow B but we claim here that it is also the equalizer of c_1, c_2: Y \longrightarrow Y \sqcup_{I}Y.

:Indeed, by construction b_1\, m = b_2\, m thus the "cokernel pair" diagram for m yields a unique morphism u': Y \sqcup_{I}Y \longrightarrow B such that b_1 = u'\, c_1,\ b_2 = u'\, c_2. Now, a map m': I'\longrightarrow Y which equalizes (c_1, c_2) also satisfies b_1\, m'= u'\, c_1 \, m'= u'\, c_2\, m'= b_2\, m', hence by the equalizer diagram for (b_1, b_2), there exists a unique map h': I'\to I such that m'= m\, h'.

:Finally, use the cokernel pair diagram (of f) with j_1 := c_1,\ j_2 := c_2,\ Z:= Y\sqcup_I Y : there exists a unique u: Y \sqcup_{X}Y \longrightarrow Y\sqcup_I Y such that c_1 = u\, i_1,\ c_2 = u\, i_2. Therefore, any map g which equalizes (i_1, i_2) also equalizes (c_1, c_2) and thus uniquely factorizes as g= m\, h' . This exactly means that (I,m) is the equalizer of (i_1, i_2).

Second definition implies the first:

  • Factorization: taking m' := f in the equalizer diagram ( m' corresponds to g ), one obtains the factorization f = m\, h .
  • Universality: let f = m'\, e' be a factorization with m' regular monomorphism, i.e. the equalizer of some pair (d_1, d_2) .

File:Equalizerd1d2.png

:Then d_1\, m'= d_2\, m'\ \Rightarrow \ d_1\, f=d_1\, m'\, e= d_2\, m'\, e= d_2\, f so that by the "cokernel pair" diagram (of f), with j_1 := d_1,\ j_2 := d_2,\ Z:= D, there exists a unique u: Y \sqcup_{X}Y \longrightarrow D such that d_1 = u\, i_1,\ d_2 = u''\, i_2.

:Now, from i_1\, m= i_2\, m (m from the equalizer of (i1, i2) diagram), one obtains d_1\, m= u\, i_1\, m = u\, i_2\, m = d_2\, m, hence by the universality in the (equalizer of (d1, d2) diagram, with f replaced by m), there exists a unique v: Im \longrightarrow I' such that m = m'\, v.

}}

Examples

In the category of sets the image of a morphism f\colon X \to Y is the inclusion from the ordinary image \{f(x) ~|~ x \in X\} to Y. In many concrete categories such as groups, abelian groups and (left- or right) modules, the image of a morphism is the image of the correspondent morphism in the category of sets.

In any normal category with a zero object and kernels and cokernels for every morphism, the image of a morphism f can be expressed as follows:

:im f = ker coker f

In an abelian category (which is in particular binormal), if f is a monomorphism then f = ker coker f, and so f = im f.

Essential Image

A related notion to image is essential image.{{Cite web |title=essential image in nLab |url=https://ncatlab.org/nlab/show/essential+image |access-date=2024-11-15 |website=ncatlab.org}}

A subcategory C \subset B of a (strict) category is said to be replete if for every x \in C, and for every isomorphism \iota: x \to y, both \iota and y belong to C.

Given a functor F \colon A \to B between categories, the smallest replete subcategory of the target n-category B containing the image of A under F.

See also

References