This is a subsection of the entry cohesive (∞,1)-topos. See there for background and context.
A cohesive (∞,1)-topos is a context of ∞-groupoids that are equipped with a geometric notion of cohesion on their collections of objects and k-morphisms, for instance topological cohesion or smooth cohesion.
While the axioms of cohesion do imply the intrinsic existence of exponentiated infinitesimal spaces, they do not admit access to an explicit synthetic notion of infinitesimal extension.
Here we consider one extra axiom on a cohesive (∞,1)-topos that does imply a good intrinsic notion of synthetic differential extension, compatible with the given notion of cohesion. We speak of differential cohesion.
In a cohesive -topos with differential cohesion there are for instance good intrinsic notions of formal smoothness and of de Rham spaces of objects.
We discuss extra structure on a cohesive (∞,1)-topos that encodes a refinement of the corresponding notion of cohesion to infinitesimal cohesion . More precisely, we consider inclusions of cohesive
-toposes that exhibit the objects of as infinitesimal cohesive neighbourhoods of objects in .
Given a cohesive -topos we say that an infinitesimal cohesive neighbourhood of is another cohesive -topos equipped with an adjoint quadruple of adjoint (∞,1)-functors of the form
where is a full and faithful and preserves finite products.
Equivalently this means that is a local geometric morphism with a further right adjoint to the right adjoint to the direct image.
Conversely we will say that data as in def. 1 equips the cohesive -topos with differential cohesion.
By the characterizaton of full and faithful adjoint (∞,1)-functors the condition on is equivalent to . Since it follows by essential uniqueness of adjoint (∞,1)-functors that also .
The inclusion into the infinitesimal neighbourhood is necessarily a morphism of (∞,1)-toposes over ∞Grpd.
as is the induced geometric morphism
Moreover is necessarily a full and faithful (∞,1)-functor.
By essential uniqueness of th global section geometric morphism: In both cases the direct image functor has as left adjoint that preserves the terminal object. Therefore
Analogously in the second case.
We shall write
so that the global section geometric moprhism of factors as
We also consider the (∞,1)-monads/comonads induced from these reflections:
the reduction modality ;
the infinitesimal shape modality ;
the infinitesimal flat modality .
The above says that these interact with the modalities of the ambient cohesion, i.e.
the shape modality ;
the flat modality ;
the sharp modality
Here the inclusion sign is to mean that the modal types of the modality on the left are included in the modal types of the modality on the right.
Let for the remainder of this section an infinitesimal neighbourhood be fixed.
From -sheaves over infinitesimal neighbourhood sites
We give a presentation of classes of infinitesimal neighbourhoods by simplicial presheaves on suitable sites.
Let be an ∞-cohesive site. We say a site
is an infinitesimal neighbourhood site of .
Let be an ∞-cohesive site and an infinitesimal neighbourhood site.
Then the (∞,1)-category of (∞,1)-sheaves on is a cohesive -topos and the restriction along exhibits it as an infinitesimal neighbourhood of the cohesive -topos over .
Moreover, restricts on representables to the (∞,1)-Yoneda embedding factoring through :
We present the (∞,1)-sheaf (∞,1)-category by the projective model structure on simplicial presheaves left Bousfield localized at the covering sieve inclusions
(as discussed at models for (∞,1)-sheaf (∞,1)-toposes).
Consider the right Kan extension of simplicial presheaves along the functor . On an object it is given by the end-expression
where in the last step we use the Yoneda reduction-form of the Yoneda lemma.
This shows that the right adjoint to is itself given by precomposition with a functor, and hence has itself a further right adjoint, which gives us a total of four adjoint functors
From this are directly induced the corresponding simplicial Quillen adjunctions on the global projective and injective model structure on simplicial presheaves
Observe that , being a left Kan extension, sends representables to representables: we have
and by Yoneda reduction (more explicitly: observing that this is equivalently the formula for left Kan extension of the non-corepresentable along the identity functor) this is
By the discussion at simplicial Quillen adjunction for the above Quillen adjunctions to descend to the Cech-local model structure on simplicial presheaves it suffices that the right adjoints preserve locally fibrant objects.
We first check that sends locally fibrant objects to locally fibrant objects.
To that end, let be a covering family in . Write for its Cech nerve, where denotes the Yoneda embedding. Recall by the definition of the ∞-cohesive site that all the fiber products of representable presheaves here are again themselves representable, hence . This means that the left adjoint preserves not only the coend and tensoring, but by the remark in the previous paragraph and the assumption that preserves pullbacks along covers we have that
By the assumption that preserves covers, this is the Cech nerve of a covering family in . Therefore for fibrant we have for all coverings in that the descent morphism
is a weak equivalence, hence that is locally fibrant.
To see that preserves locally fibrant objects, we apply the analogous reasoning after observing that its left adjoint preserves all limits and colimits of simplicial presheaves (as these are computed objectwise) and by observing that for a covering family in we have that its image under is its image under , by the Yoneda lemma:
and using that preserves covers by assumption.
Therefore is a left and right local Quillen functor with left local Quillen adjoint and right local Quillen adjoint .
It follows that is given by the left derived functor of restriction along , and is given by the right derived functor of restriction along .
Finally to see that also preserves locally fibrant objects by the same reasoning as above, notice that for every covering family in and every morphism in we may find a covering of such that we find commuting diagrams on the left of
because by adjunction these correspond to commuting diagrams as indicated on the right, which exist by definition of coverage on and lift through by assumption on .
This implies that is a generalized cover in the terminology at model structure on simplicial presheaves, which by the discussion there implies that the corresponding Cech nerve equivalent to the sieve inclusion is a weak equivalence.
This establishes the quadruple of adjoint (∞,1)-functors as claimed.
It remains to see that is full and faithful. For that notice the general fact that left Kan extension (see the properties discussed there) along a full and faithful functor satisfies . It remains to observe that since is not only right but also left Quillen by the above, we have that applied to a cofibrant object is already the derived functor of the composite.
Relation to infinitesimal cohesion
We discuss how differential cohesion in the sense of def. 1 relates to infinitesimal cohesion.
Given differential cohesion, def. 1,
define operations and by
Hence makes a homotopy pushout square
and makes a homotopy pullback square
We call the relative shape modality and the relative flat modality.
The relative shape and flat modalities of def. 3
form an adjoint pair ;
whose (co-)modal types are precisely the properly infinitesimal types, hence those for which is an equivalence;
preserves the terminal object.
It follows that when has a further right adjoint with equivalent modal types containing the codiscrete types, then this defines a level
hence an intermediate subtopos which is infinitesimally cohesive.
This happens notably for the model of formal smooth ∞-groupoids and all its variants such as formal complex analytic ∞-groupoids etc. But in this case does not provide Aufhebung for .
Structures in a differential cohesive -topos
We discuss structures that are canonically present in a cohesive -topos equipped with differential cohesion. These structures parallel the structures in a general cohesive (∞,1)-topos.
Infinitesimal paths and de Rham spaces
In the presence of differential cohesion there is an infinitesimal analog of the geometric paths ∞-groupoids.
Infinitesimal path -groupoid
Define the adjoint triple of adjoint (∞,1)-functors corresponding to the adjoint quadruple :
We say that
An object in the full sub--category
For we say that
There is a canonical natural transformation
that factors the finite path inclusion through the infinitesimal one
This is the formula for the unit of the composite adjunction :
Notice that for any morphism in any (∞,1)-topos , there is the corresponding base change geometric morphism between the over-(∞,1)-toposes
For any object write
for the base change geometric morphism induced by the constant infinitesimal path inclusion , def. 4.
For we call as well as its pullback (depending on context) the jet bundle of .
Formally smooth/étale/unramified morphisms
We say an object is formally smooth if the constant infinitesimal path inclusion, , def. 4,
is an effective epimorphism.
In this form this is the evident -categorical analog of the conditions as they appear for instance in (SimpsonTeleman, page 7).
The canonical morphism is the composite
By the condition that is a full and faithful (∞,1)-functor the second morphism here in an equivalence, as indicated, and hence the component of the composite on being an effective epimorphism is equivalent to the component being an effective epimorphism.
Therefore we have the following more general definition.
Even more generally we can formulate formal smoothness in :
A morphism in is formally étale if it is -closed, hence if its -unit naturality square
is an (∞,1)-pullback.
This is again given by the fact that by definition and that is fully faithful, so that
The collection of formally étale morphisms in , def. 7, is closed under the following operations.
Every equivalence is formally étale.
The composite of two formally étale morphisms is itself formally étale.
is a diagram such that and are formally étale, then also is formally étale.
Any retract of a formally étale morphisms is itself formally étale.
The (∞,1)-pullback of a formally étale morphisms is formally étale if the pullback is preserved by .
The statements about closure under composition and pullback appears as(KontsevichRosenberg, prop. 5.4, prop. 5.6). Notice that the extra assumption that preserves the pullback is implicit in their setup, by remark 8.
The first statement follows since -pullbacks are well defined up to quivalence.
The second two statements follow by the pasting law for (∞,1)-pullbacks: let and be two morphisms and consider the pasting diagram
If and are formally étale then both small squares are pullback squares. Then the pasting law says that so is the outer rectangle and hence is formally étale. Similarly, if and are formally étale then the right square and the total reactangle are pullbacks, so the pasting law says that also the left square is a pullback and so also is formally étale.
For the fourth claim, let be a retract in the arrow (∞,1)-category . By applying the natural transformation we obtain a retract
in the category of squares . We claim that generally, if the middle piece in a retract in is an (∞,1)-pullback square, then so is its retract sqare. This implies the fourth claim.
To see this, we use that
(∞,1)-limits are computed by homotopy limits in any presentable (∞,1)-category presenting ;
homotopy limits in may be computed by the left and right adjoints provided by the derivator associated to .
From this the claim follows as described in detail at retract in the section retracts of diagrams .
For the last claim, consider an (∞,1)-pullback diagram
where is formally étale.
Applying the natural transformation to this yields a square of squares. Two sides of this are the pasting composite
and the other two sides are the pasting composite
Counting left to right and top to bottom, we have that
the first square is a pullback by assumption that preserves the given pullback;
the second square is a pullback, since is formally étale.
the total top rectangle is therefore a pullback, by the pasting law;
the fourth square is a pullback since is right adjoint and so also preserves pullbacks;
also the total bottom rectangle is a pullback, since it is equal to the top total rectangle;
therefore finally the third square is a pullback, by the other clause of the pasting law. Hence is formally étale.
In order to interpret the notion of formal smoothness, we turn now to the discussion of infinitesimal reduction.
The operation is an idempotent projection of onto the image of
By definition of infinitesimal neighbourhood we have that is a full and faithful (∞,1)-functor. It follows that and hence
For every , we have that is formally smooth according to def. 6.
A set of objects is said to exhibit the differential structure or
exhibit the infinitesimal thickening if the localization
of at the morphisms of the form is exhibited by the infinitesimal shape modality .
For more see at Lie differentiation.
We discuss how in differential cohesion every object canonically induces its étale (∞,1)-topos .
For any object in a differential cohesive -topos, we formulate
the (∞,1)-topos denoted or of (∞,1)-sheaves over , or rather of formally étale maps into ;
the structure (∞,1)-sheaf of .
The resulting structure is essentially that discussed (Lurie, Structured Spaces) if we regard equipped with its formally étale morphisms, def. 8, as a (large) geometry for structured (∞,1)-toposes.
One way to motivate this is to consider structure sheaves of flat differential forms. To that end, let a differential cohesive ∞-group with de Rham coefficient object and for any differential homotopy type, the product projection
regarded as an object of the slice (∞,1)-topos almost qualifies as a “bundle of flat -valued differential forms” over : for an cover (a 1-epimorphism) regarded in , a -plot of this product projection is a -plot of together with a flat -valued de Rham cocycle on .
This is indeed what the sections of a corresponding bundle of differential forms over are supposed to look like – but only if is sufficiently spread out over , hence sufficiently étale. Because, on the extreme, if is the point (the terminal object), then there should be no non-trivial section of differential forms relative to over , but the above product projection instead reproduces all the sections of .
In order to obtain the correct cotangent-like bundle from the product with the de Rham coefficient object, it needs to be restricted to plots out of suficiently étale maps into . In order to correctly test differential form data, “suitable” here should be “formally”, namely infinitesimally. Hence the restriction should be along the full inclusion
of the formally étale maps of def. 8 into . Since on formally étale covers the sections should be those given by , one finds that the corresponding “cotangent bundle” must be the coreflection along this inclusion. The following proposition establishes that this coreflection indeed exists.
For any object, write
for the full sub-(∞,1)-category of the slice (∞,1)-topos over on those maps into which are formally étale, def. 8.
We also write or for .
The inclusion of def. 10 is both reflective as well as coreflective, hence it fits into an adjoint triple of the form
By the general discussion at reflective factorization system, the reflection is given by sending a morphism to and the reflection unit is the left horizontal morphism in
Therefore , being a reflective subcategory of a locally presentable (∞,1)-category, is (as discussed there) itself locally presentable. Hence by the adjoint (∞,1)-functor theorem it is now sufficient to show that the inclusion preserves all small (∞,1)-colimits in order to conclude that it also has a right adjoint (∞,1)-functor.
So consider any diagram (∞,1)-functor out of a small (∞,1)-category. Since the inclusion of is full, it is sufficient to show that the -colimit over this diagram taken in lands again in in order to have that -colimits are preserved by the inclusion. Moreover, colimits in a slice of are computed in itself (this is discussed at slice category - Colimits).
Therefore we are reduced to showing that the square
is an (∞,1)-pullback square. But since is a left adjoint it commutes with the -colimit on objects and hence this diagram is equivalent to
This diagram is now indeed an (∞,1)-pullback by the fact that we have universal colimits in the (∞,1)-topos , hence that on the left the component for each is the (∞,1)-pullback of , by assumption that we are taking an -colimit over formally étale morphisms.
For the case that in prop. 8, then the proof there shows that the étalification operation over the point is just :
Indeed, for any then is a formally étale morphism since
is a homotopy pullback.
The -category is an (∞,1)-topos and the canonical inclusion into is a geometric embedding.
By prop. 8 the inclusion is reflective with reflector given by the factorization system. Since is a right adjoint and hence in particular preserves (∞,1)-pullbacks, the -equivalences are stable under pullbacks. By the discussion at stable factorization system this is the case precisely if the corresponding reflector preserves finite (∞,1)-limits. Hence the embedding is a geometric embedding which exhibits a sub-(∞,1)-topos inclusion.
For we speak of
also as the (petit) (∞,1)-topos of or the étale (∞,1)-topos of .
for the composite (∞,1)-functor that sends any to the etalification, prop. 8, of the projection .
We call the structure sheaf of .
This is the analog of (Lurie, Structured Spaces, prop. 2.2.11).
We need to check that preserves finite (∞,1)-limits and formally étale covers (where covers here in the canonical topology on the given toposes are 1-epimorphisms). The first statement follows since is right adjoint to the forgetful functor
For the second statement, let be any 1-epimorphism which is also a formally étale. We need to show that also is a 1-epimorphism. By the discussion at effective epimorphism in an (∞,1)-category for this it is sufficient that the 0-truncation is an epimorphism in the underlying sheaf topos, hence that every generalized element of has a lift to after refinement along a cover.
By the fact that is right adjoint to the inclusion, by construction, this means that it is sufficient to show that given a diagram in of the form
with formally étale, this can be completed to a diagram of the form
with a formally étale 1-epimorphism. But since both 1-epimorphisms as well as formally étale morphisms are stable under (∞,1)-pullback we can take .
Let be an ∞-group and write for the corresponding de Rham coefficient object. Then
we may call the -valued flat cotangent sheaf of .
For an object in a differentially cohesive -topos, then its petit structured -topos , according to def. 11, is locally ∞-connected.
We need to check that the composite
preserves (∞,1)-limits, so that it has a further left adjoint. Here is the reflector from prop. 8. Inspection shows that this composite sends an object to :
By the discussion at slice (∞,1)-category – Limits and colimits an (∞,1)-limit in the slice is computed as an (∞,1)-limit in of the diagram with the slice cocone adjoined. By right adjointness of the inclusion the same is then true for .
Now for a diagram, it is taken to the diagram in and so its -limit is computed in over the diagram locally of the form
Since -limits commute with each other this limit is the product of
(over the co-coned diagram constant on ).
For the first of these, since the infinitesimal shape modality is in particular a right adjoint (with left adjoint the reduction modality), and since is also right adjoint by cohesion, we have a natural equivalence
For the second, the -limit over an -category of a functor constant on is
where the last line follows since has a terminal object and hence contractible geometric realization.
In conclusion this shows that -limits are preserved by .
Since the inclusion of the point into the interval is an op-final (∞,1)-functor we have (by this proposition) an equivalence of over-(∞,1)-categories
Since is formally étale by assumption and since formally étale morphisms are closed under composition, this restricts to an equivalence .
For the equivalence of structure sheaves it is sufficient to show for each coefficient an equivalence
in . But by definition (11) and similarly for and since is right adjoint to the inclusion we have
Sheaves of (quasi-coherent) modules
We discuss the abstract formulation of sheaves of modules and of quasicoherent sheaves on petit -toposes in differential cohesion.
under construction – check
For an object and its -structured (infinity,1)-topos, according to def. 11, consider the composite functor
then an -module on is an extension of by a limit-preserving functor through
In particular is canonically a module over itself by setting .
This is a slight abstraction of the definition in (Lurie QC, section 2.3). See at quasicoherent sheaf – In higher geometry – As extension of the structure sheaf.
For and with as in def. 11, write
for the morphism in which is the -adjunct of the counit of the -coreflection of def. 11.
This we call the Liouville-Poincaré -cocycle on .
Consider the model of differential cohesion given by SynthDiff∞Grpd. Write for the abstract sheaf of differential 1-forms.
Then for a smooth manifold, we have that
is the cotangent bundle
of the manifold: because for an open subset of the manifold regarded as an object of , a section of is equivalently a map in , which by the -adjunction is a map in which finally is equivalently a map in hence an element in .
So the Liouville-Poincaré -cocycle according to 13 is a differential 1-form
on the total space of the cotangent bundle. For
a section of the cotangent bundle, the pullback form on is the composite
hence the adjunct
hence by definition
hence the adjunct
hence the original . This is the defining property which identifies as the traditional Liouville-Poincaré 1-form.
Manifolds and étale groupoids
An ordinary topological/Lie étale groupoid is one whose source/target map is an étale map. We consider now a notion that can be formulated in the presence of infinitesimal cohesion which generalizes this.
Let now be any object
(For instance let be a canonical line object that exhibits the cohesion of in the sense of structures in a differential infinity-topos – A1 homotopy / The continuum and take .)
A -manifold is an object such that there exists a -cover , namely a correspondence from to
such that both morphisms are formally étale morphisms and such that is in addition an effective epimorphism.
See also at smooth manifold – general abstract geometric formulation
We discuss how each manifold in differential cohesion as in def. 15 is associated a canonical frame bundle classified by a morphism .
For any object in differential cohesion, its infinitesimal disk bundle is the homotopy pullback
of the unit of its infinitesimal shape modality along itself.
More generally, given a filtration of differential cohesion by orders of infinitesimals, remark 4, then the order- infinitesimal disk bundle is the homotopy pullback in
If is a formally étale morphism, def. 8, then
By the definition of formal étaleness and using the pasting law we have an equivalence of pasting diagrams of homotopy pullbacks of the following form:
For an object, a framing on is a trivialization of its infinitesimal disk bundle, def. 16, i.e. an object – the typical infinitesimal disk or formal disk – and a (chosen) equivalence
This class of examples of framings is important:
Every differentially cohesive ∞-group is canonically framed (def. 17) such that the horizontal map in def. 16 is given by the left action of on its infinitesimal disk at the neutral element:
By the discussion at Mayer-Vietoris sequence in the section Over an ∞-group and using that the infinitesimal shape modality preserves group structure, the defining homotopy pullback of is equivalent to the pasting of pullback diagrams of the form
where the right square is the defining pullback for the infinitesimal disk . For the left square we find by this proposition that and that the top horizontal morphism is as claimed.
By lemma 1 it follows that:
For a framed object, def. 17, let be a -manifold, def. 15. Then the infinitesimal disk bundle, def. 16, of canonically trivializes over any -cover , i.e. there is a homotopy pullback of the form
This exhibits as a -fiber ∞-bundle.
We discuss the formalization of G-structures and integrability of G-structures in differential cohesion
Let be framed, def. 17, let be an ∞-group and a homomorphism to the general linear group of , def. 18, hence
a morphism between the deloopings.
For a -manifold, def. 15, a G-structure on is a lift of the structure group of its frame bundle, def. 19, to , hence a diagram
hence a morphism
in the slice (∞,1)-topos.
In fact is the moduli ∞-stack of such -structures.
The double slice is the (∞,1)-category of such -structures.
If is framed, def. 17, then it carries the trivial -structure, which we denote by
For framed, def. 17, and a -manifold, def. 15, then a -structure on , def. 20, is integrable (or locally flat) if there exists a -cover
such that the correspondence of frame bundles induced via remark 19
(a diagram in ) extends to a sliced correspondence between and the trivial -structure on , example 5, hence to a diagram in of the form
On the other hand, is called infinitesimally integrable (or torsion-free) if such an extension exists (only) after restriction to all infinitesimal disks in and , hence after composition with the counit
of the relative flat modality, def. 3 (using that by prop. 4 this is also formally étale and hence induces map of frame bundles):
Flat -connections and infinitesimal local systems
We discuss the intrinsic flat cohomology in an infinitesimal neighbourhood.
For we say that
(where is given by def. 4) is the infinitesimal flat cohomology of with coefficient in .
In traditional contexts this is also called crystalline cohomology or just de Rham cohomology . Since we already have an intrinsic notion of de Rham cohomology in any cohesive (∞,1)-topos, which is similar to but may slightly differ from infinitesimal flat differential cohomology, we shall say synthetic de Rham cohomology for the notion of def. 22 if we wish to honor traditional terminology. In this case we shall write
By the above observation we have canonical morphisms
The objects on the left are principal ∞-bundles equipped with flat ∞-connection . The first morphism forgets their higher parallel transport along finite volumes and just remembers the parallel transport along infinitesimal volumes. The last morphism finally forgets also this connection information.
For an abelian ∞-group object we say that the de Rham theorem for -coefficients holds in if for all the infinitesimal path inclusion
is an equivalence in -cohomology, hence if for all we have that
is an isomorphism.
If we follow the notation of note 1 and moreover write for the intrinsic geometric realization, then this becomes
where on the right we have ordinary cohomology in Top (for instance realized as singular cohomology) with coefficients in the discrete group underlying the cohesive group .
In certain contexts of infinitesimal neighbourhoods of cohesive -toposes the de Rham theorem in this form has been considered in (SimpsonTeleman).
Formal cohesive -groupoids
Recall that a groupoid object in an (infinity,1)-category is equivalently an 1-epimorphism , thought of as exhibiting an atlas for the groupoid .
Now an -Lie algebroid is supposed to be an -groupoid which is only infinitesimally extended over its base space . Hence:
A groupoid object is infinitesimal if under the reduction modality (equivalently under the infinitesimal shape modality ) the atlas becomes an equivalence: .
For example the tangent -Lie algebroid of any is the unit of the infinitesimal shape modality.
It follows that every such -Lie algebroid canonically maps to the tangent -Lie algebroid of – the anchor map. The naturality square of the unit exhibits the morphism:
The discussion at synthetic differential ∞-groupoid – Lie differentiation immediately generalizes to produce a concept of Lie differentiation in any differentially cohesive context. This Lie differentiation is just the flat modality of the differential cohesion but regarded as cohesive over its induced infinitesimal cohesion. As such, there is a left adjoint to Lie differentiation, given by the corresponding shape modality. However, the substance of Lie theory here will be to restrict this adjunction to geometric ∞-stacks. On the geometric -stacks the Lie differentiation via passage to inffinitesimal cohesion will yield actual -algebras, but some structure is required to make the formal Lie integration of these lang indeed in geometric ∞-stacks.
For any (∞,1)-site the construction of the tangent (∞,1)-category provides a canonical infinitesimal thickening of :
where the -functor pair on the right forms a -relative (∞,1)-adjunction. The composite is the cotangent complex functor for and is fiberwise the canonical map out of the stabilization.
The image of is contained in that of . Therefore we may restrict the -adjunction on the right to the full sub-(∞,1)-category of on thise objects in the image of . This yields an infinitesimal neighbourhood of (∞,1)-sites
See at differential cohesion and idelic structure.
graded differential cohesion
The material discussed here corresponds to the most part to sections 3.5 and 3.10 of
For references on the general notion of cohesive (∞,1)-topos, see there.
The following literature is related to or subsumes by the discussion here.
Something analogous to the notion of ∞-connected site and the fundamental ∞-groupoid in a locally ∞-connected (∞,1)-topos is the content of section 2.16. of
The infinitesimal path ∞-groupoid adjunction is essentially discussed in section 3 there.
The characterization of infinitesimal extensions and formal smoothness by adjoint functors (in 1-category theory) is considered in
in the context of Q-categories .
The notion of forming petit -toposes of étale objects over a given object appears in