nLab comprehensive factorization system

Idea

Every functor $F:C\to D$ factors as a final functor followed by a discrete fibration.

Statement

Theorem

Let $E$ be the class of final functors and $M$ be the class of discrete fibrations. Then $(E,M)$ is an orthogonal factorization system of Cat, called the comprehensive factorization system.

Proof

Let $F:C\to D$ be a functor. Define $K:D\to Set$ as the left Kan extension of the constant presheaf $C\to Set$ at the singleton along $F$. Explicitly, $K(d)$ is the set of connected components of $F/d$. Let $E=\int K$, so an object of $E$ is an ordered pair $(d, [\alpha:Fc\to d])$ where $[\alpha]$ denotes the connected component of $(c,\alpha)$. Then it is not hard to verify that $e:C\to E$ mapping $c\mapsto (Fc,[id_{fc}])$ is final, the canonical $m:E\to D$ is a discrete fibration, and $F=me$.

Now we show that $E$ and $M$ are replete subcategories of $Cat$. Clearly they include all isomorphisms.

If functors $F:C\to D$ and $G: D\to E$ are final, then we show that $G\circ F$ is final. For $e\in E$, there is an element $(d,\alpha:e\to Gd)$ of $e/G$, and thence an element $(c,\beta:d\to Fc)$ of $d/F$, so we obtain an element $(c, e \stackrel{\alpha}{\to} Gd \stackrel{G\beta}{\to} GFc)$ of $e/GF$. Now we must show that any two elements $(c,\gamma:e\to GFc),(c',\gamma':e\to GFc')$ are connected. Since $G$ is final, elements $(Fc,\gamma)$ and $(Fc',\gamma')$ of $e/G$ are connected. It suffices to consider the case of a zig-zag of length one: a morphism $f:Fc\to Fc'$ such that

By finality of $F$, the elements $(c,id:Fc\to Fc)$ and $(c', f:Fc\to Fc')$ of $Fc/F$ are connected. A zig-zag path between them, by precomposition with $\gamma$, becomes a zig-zag path between $(c,\gamma)$ and $(c',\gamma')$. So $G\circ F$ is final.

The proof that discrete fibrations form a subcategory is omitted.

Now we must show that the lifting problem

has a unique solution $h$ when $e\in E$ and $m\in M$.

We prove uniqueness first. For $b\in B$, let $(a,\alpha:b\to e(a))\in b/e$. Then $h(\alpha)$ must be the unique lifting of $g(\alpha)$, and $h(b)$ the domain of this lifting, proving uniqueness of $h$ on objects. For $\beta:b\to b'$ in $B$, $h(\beta)$ must be the unique lifting of $g(\beta)$, so $h$ is unique (if it exists).

Now we must show that this $h$ is well-defined, functorial, and a solution to the lifting problem. If $(a',\alpha':b\to e(a'))$ is another element of $b/e$, then WLOG let $u:a\to a'$ such that

Lifting this diagram, we see that $g(\alpha)$ and $g(\alpha')$ must lift to morphisms with identical domain, so $h$ is well-defined on objects.

For $\beta:b\to b'$ in $B$, let $\alpha:b'\in e(a)$, and by the diagram

we see that $g(\beta)$ and $g(\alpha\circ\beta)$ must lift to morphisms with identical domains, so $h(\beta)$ has domain $h(b)$.

Functoriality now follows easily from uniqueness of lifting for a discrete fibration, and it is not hard to show that $h$ is a solution to the lifting problem.

Dually, there is an orthogonal factorisation system $(E,M)$ on $Cat$ for which $E$ is the class of initial functors and $M$ is the class of discrete opfibrations.

Remark on the terminology

Under the functorial reformulation of the axiom of comprehension by Lawvere (1970) the comprehensive factorization can be viewed as a generalization of the epi-mono factorization of a function $f:X\to Y$ occurring in the context of set-theoretic comprehension: the best approximation of $f$ by a property in $2^Y$ (i.e. $\chi_{im f}$) has extension $im f\hookrightarrow Y$. Hence (in the notation of the above proof) the discrete fibration $m:\int K\to D$ given by the factorization $F=m{}e$ can be viewed as the “extension” of the approximation of $F$ by the “property” $K$ in $Set^{D}$.

More generally, instances of the comprehension scheme correspond to factorization systems (cf. Berger-Kaufmann).

Related entries

• comprehension scheme

References

• Ross Street, R. F. C. Walters, The Comprehensive Factorization of a Functor, Bulletin of the American Mathematical Society 79(5): 936-941 (bams:1183534973, doi:10.1090/S0002-9904-1973-13268-9)

Note there is a mistake in the proof of the main theorem of the paper above, as noted on page 74 of:

• Max Kelly, Basic concepts of enriched category theory, London Math. Soc. Lec. Note Series 64, Cambridge Univ. Press (1982), Reprints in Theory and Applications of Categories 10 (2005) 1-136 [ISBN:9780521287029, tac:tr10, pdf]

• Lawvere, F. W. Equality in hyperdoctrines and comprehension schema as an adjoint functor. In A. Heller, ed., Proc. New York Symp. on Applications of Categorical Algebra, pp. 1–14. AMS, 1970. (pdf)

• Fosco Loregian, Emily Riehl, Categorical Notions of Fibration, Expositiones Mathematicae 38, 2020. (arXiv:1806.06129, doi:10.1016/j.exmath.2019.02.004)

• Clemens Berger, Ralph M. Kaufmann, Comprehensive Factorization Systems, Tbilisi Math. J. 10 (2017), 255-277 (doi:10.1515/tmj-2017-0112)

• Paolo Perrone, Walter Tholen, Kan extensions are partial colimits, Kan Extensions are Partial Colimits. Applied Categorical Structures 30, 685–753 (2022). (arXiv:2101.04531. doi:10.1007/s10485-021-09671-9)

Internal comprehensive factorisations (and torsors) are considered in:

• Ross Street, Dominic Verity, The comprehensive factorization and torsors, TAC 23 (2010) 42-75 [tac:23-03]

A generalisation to a bicategorical factorization system on $CAT$ is considered in:

• Ross Street, Variation on a Comprehensive Theme, TAC 37 (2021) 964-978 [tac:37-29]