Paths and cylinders
Limits and colimits
limits and colimits
limit and colimit
limits and colimits by example
commutativity of limits and colimits
connected limit, wide pullback
preserved limit, reflected limit, created limit
product, fiber product, base change, coproduct, pullback, pushout, cobase change, equalizer, coequalizer, join, meet, terminal object, initial object, direct product, direct sum
end and coend
Special and general types
Mayer-Vietoris sequence is the term for the fiber sequence – or often for the corresponding long exact sequence of homotopy groups – induced from an (∞,1)-pullback (or for a homotopy pullback presenting it).
Let be an (∞,1)-category with finite (∞,1)-limits and let be pointed objects and
be any two morphisms with common codomain preserving the base points. Let be the (∞,1)-pullback
The corresponding Mayer-Vietoris sequence is the fiber sequence of the induced morphism . Often the term is used (only) for the corresponding long exact sequence of homotopy groups.
Let be a presentable (∞,1)-category.
Then is equivalently given by the (∞,1)-pullback
where the right vertical morphism is the diagonal.
Moreover, the homotopy fiber of is the loop space object .
See also at homotopy pullback this corollary.
The first statement one checks for instance by choosing a presentation by a combinatorial model category and then proceeding as below in the discussion Presentation by fibrant objects. Then by the pasting law for -pullbacks it follows that with the left square in
an -pullback, so is the total outer rectangle. But again by the first statement, this is equivalent to the -pullback
which is the defining pullback for the loop space object.
Therefore the Mayer-Vietoris homotopy fiber sequence is of the form
For ∞Grpd Top, this point of view is amplified in (Dyer-Roitberg 80).
The corresponding long exact sequence of homotopy groups is of the form
This is what has historically been the definition of Mayer-Vietories sequences (Eckmann-Hilton 64).
Presentation by fibrant objects
Suppose that the (∞,1)-category is presented by a category of fibrant objects (for instance the subcategory on the fibrant objects of a model category).
Then the -pullback is presented by a homotopy pullback, and by the factorization lemma, this is given by the ordinary limit
where is a path object for . This limit coincides, up to isomorphism, with the pullback
This implies in particular that the homotopy fiber of is the loop space object , being the fiber of the path space object projection.
Over an -group
We consider now the case where carries the structure of an ∞-group (or just a grouplike H-space object) in a presentable (∞,1)-category or locally Cartesian closed (∞,1)-category .
In this case (as discussed in a moment), we have an (∞,1)-pullback
where the bottom horizontal morphism is the composite
of a morphism that sends the second argument to its inverse with the group composition operation.
It then follows by the pasting law and prop. 1 that in this case the morphism in the Mayer-Vietoris sequence is itself the homotopy fiber of , hence that we have a long homotopy fiber sequence of the form
First consider two more concrete special cases.
Let be a small site and let be the (∞,1)-category of (∞,1)-sheaves on .
This is presented by the projective model structure on simplicial presheaves
As discussed there, the Dold-Kan correspondence prolongs to a Quillen adjunction on presheaves whose right adjoint is
Let then be an object with a presentation in in the image of this . We write also for this presentation, and hence for some presheaf of chain complexes .
We claim now that such satisfies the above assumption.
To see this, first notice that the evident morphism is degreewise an epimorphism, hence it is a fibration in , and since is right Quillen, so is the corresponding morphism in .
Therefore the ordinary pullback of presheaves of chain complexes
is a homotopy pullback in , as is the ordinary pullback of simplicial presheaves
Since ∞-stackification preserves finite (∞,1)-limits, this presents an (∞,1)-pullback also in .
Let be an (∞,1)-topos with a 1-site of definition (a 1-localic (∞,1)-topos).
Then (as discussed there) every ∞-group object in has a presentation by a presheaf of simplicial groups
We claim that the canonical morphism is objectwise a Kan fibration and hence a fibration in the projective model structure on simplicial presheaves.
Let be any test object. A diagram
corresponds to a -cell together with a choice of decomposition of the th horn as a difference
Since itself is a Kan complex (being a simplicial group, as discussed there) there is a filler of the horn . Define then
Since all the face maps are group homomorphisms, this is indeed a filler of :
Moreover, by construction, is a filler in
Since therefore is a projective fibration, it follows as before that the ordinary pullback
is a homotopy pullback.
For an ∞-group object as above, the (∞,1)-pullback is equivalently given by the -pullback
By prop. 1 the object is the -pullback in
By the pasting law this is equivalently given by the composite pullback of
Here the composite bottom morphism is .
Summing this up:
For an (∞,1)-sheaf (∞,1)-topos, an ∞-group-object in and and two morphisms, then there is a long homotopy fiber sequence of the form
For an (∞,1)-site of definition, there is a reflection
of into an (∞,1)-category of (∞,1)-presheaves.
By prop. 2 the statement holds in . Since embedding and reflection both preserve finite (∞,1)-limits, it hence also holds in .
Still more generally and more simply:
Let be a locally Cartesian closed (∞,1)-category. Let be an ∞-group object (or just a grouplike H-space-object). Then for any morphism we have a homotopy pullback square of the form
By this discussion-category#InternalLogic) we may use homotopy type theory reasoning. Starting out with the discussion at homotopy pullback – In homotopy type theory we obtain
where the second but last step consists of observing a contractible based path space object (see the discussion at factorization lemma).
Let be a locally Cartesian closed (∞,1)-category. Let be an ∞-group object (or just a grouplike H-space-object).
Then for and two morphisms, there is a Mayer-Vietoris-type homotopy fiber sequence
Use prop. 4 with being the canonical point, i.e. the inclusion of the neutral element to find the homotopy pullback
Then use the pasting law as above.
(Co)Homology of a cover
A special case of the general Mayer-Vietoris sequence, corollary 1 – which historically was the first case considered – applies to the cohomology/homology of a topological space equipped with an open cover .
Being a cover means (see effective epimorphism in an (∞,1)-category) that there is a homotopy pushout diagram of the form
in the (∞,1)-topos ∞Grpd/Top.
When this is presented by the standard model structure on simplicial sets or in terms of CW-complexes by the model structure on topological spaces, it is given by an ordinary pushout.
Let then be some coefficient object, for instance an Eilenberg-MacLane object (Eilenberg-MacLane space ) for the definition of ordinary singular cohomology with coefficients in an abelian group .
Then applying the derived hom space functor yields the (∞,1)-pullback diagram
to which we can apply the homotopical Mayer-Vietoris sequence.
Notice that (as discussed in detail at cohomology) the homotopy groups of the ∞-groupoid are the cohomology groups of with coefficients in
By the above general properties the above homotopy pullback is equivalent to
being a fiber sequence. The corresponding long exact sequence in cohomology (as discussed above) is what is traditionally called the Mayer-Vietoris sequence of the cover of by and in -cohomology.
By duality (see universal coefficient theorem) an analogous statement holds for the homology of , and .
An original reference is
A more modern review that emphasizes the role of homotopy fiber sequences is in
- Eldon Dyer, Joseph Roitberg, Note on sequence of Mayer-Vietoris type, Proceedings of the AMS, volume 80, number 4 (1980) (pdf)
Discussion in the context of stable model categories includes
- Peter May, lemma 5.7 of The additivity of traces in triangulated categories, Adv. Math., 163(1):34-73, 2001 (pdf)
Discussion in the context of homotopy type theory includes
- E Cavallo et al, Exactness of the Mayer-Vietoris Sequence in Homotopy Type Theory (pdf)