nLab
homotopy 2-type

Idea

A homotopy 2-type is a space where we consider its properties only up to the 2nd homotopy group π 2.

Definition

A continuous map XY is a homotopy 2-equivalence if it induces isomorphisms on π i for i=0,1,2 at each basepoint. Two spaces share the same homotopy 2-type if they are linked by a zig-zag chain of homotopy 2-equivalences.

For any nice space X, you can kill its homotopy groups in higher dimensions by attaching cells, thus constructing a new space Y so that the inclusion of X into Y is a homotopy 2-equivalence; up to (weak) homotopy equivalence, the result is the same for any space with the same homotopy 2-type. Accordingly, a homotopy 2-type may alternatively be defined as a space with trivial π i for i>2, or as the unique (weak) homotopy type of such a space, or as its fundamental -groupoid (which will be a 2-groupoid).

See the general discussion in homotopy n-type.

Classification

Homotopy 2-types can be classified by various different types of algebraic data.

Homotopy 2-types as crossed modules

Homotopy 2-types can be classified up to weak homotopy type by crossed modules of groupoids. These are the 2-truncated versions of crossed complexes. So such a C consists of a morphism

δ:C 2C 1\delta: C_2 \to C_1

of groupoids with object set C 0 such that C 2 is totally disconnected, i.e. is a family of groups C 2(x),xC 0. Further the groupoid C 1 operates on this family of groups so that (using right operations) if a:xy in C 1 and uC 2(x) then u aC(y); and the usual rules for an operation are satisfied, namely (uv) a=u ava, u 1=u, (u a) b=u ab when these are defined. Further the two basic crossed module rules hold:

CM1) δ(u a)=a 1(δu)a;

CM2) v 1uv=u δv;

for all aC 1,u,vC 2 when the rules make sense.

Such a crossed module may be extended to a crossed complex sk 2C by adding trivial elements in dimensions higher than 2. Hence there is a simplicial nerve N ΔC which in dimension n is

Crs(Π(Δ * n),sk 2C).Crs(\Pi (\Delta^n_*), sk^2 C).

The geometric realisation of this is the classifying space BC. Its first and second homotopy groups at xC 0 are the cokernel and kernel of δ:C 2(x)C 1(x,x). It components are those of the groupoid C 1. All other homotopy groups are trivial.

If X is a CW-complex then there is a bijection of homotopy classes

[X,BC][ΠX *,sk 2C],[X,BC] \cong [\Pi X_*, sk^2 C],

and hence there is a map XB(cotr 2ΠX *) inducing isomorphisms of homotopy groups in dimensions 1 and 2.

Here the cotruncation cotr nD of a general crossed complex D agree with D up to dimension (n1), is Cokδ n+1 in dimension n, and is trivial in higher dimensions.

It is in this sense that crossed modules of groupoids classify weak homotopy 2-types.

The category Crs 2 of such crossed modules of groupoids is equivalent to that of strict 2-groupoids. Further, Crs 2 is monoidal closed:

Crs 2(CD,E)Crs 2(C,CRS 2(D,E))Crs^2(C \otimes D, E) \cong Crs^2(C, CRS^2(D,E))

and with a unit interval object I so that (left) homotopies are determined as morphisms Crs 2(ID,E) or as elements of CRS 2(D,E) 1.

Homotopy 2-types as simplicial group(oid)s

As a crossed module give rise to an internal groupoid in the category of groups (or groupoids), we can take the nerve of that structure and get a simplicial group (or simplicially enriched groupoid). From a simplicial group(oid), G, one can define a simplicial set called the classifying space W¯G of the simplicial group, G, for which construction see simplicial group. We thus can start with a crossed module C form a simplicial group and then take W¯ of that to get another model of C.

Homotopy 2-types as 2-groupoids