1. A note on strong homology of inverse systems
- Author
-
Sibe Mardešić
- Subjects
Steenrod homology ,Pure mathematics ,Functor ,Existential quantification ,topological spaces ,strong homology ,Inverse ,Homology (mathematics) ,Combinatorics ,symbols.namesake ,Weierstrass factorization theorem ,Natural transformation ,symbols ,Singular homology ,Mathematics - Abstract
Ju. T. Lisica and the author have defined in [4] strong homology groups HP(X;G) of inverse systems of spaces X={XX, pxx-, A) over directed cofinite sets A (every element AeA has only finitelymany predecessors). It was shown in [5] that these groups are functors on the coherent prohomotopy category CPHTop, introduced in [2] and [3]. The notion of strong or Steenrod homology H£(X; G) of an arbitrary space X was then defined [1], [6] and shown to be a functor on the strong shape category SSh [2], [3]. The procedure consisted in choosing a cofiniteANR-resolution p:X->X of Z([7], [8], [9]) and of defining HSV{X; G) as HV{X;G). That the group HSP{X; G) does not depend on the choice of the resolution is a consequence of the following factorizationtheorem ([3], Theorem II.2.3). If p:X->Xi$ a resolution and f:X-*Y is a coherent map into a cofiniteANR-system, then there exists a unique coherent homotopy class of coherent maps g: X―> Y such that gp and / are coherently homotopic. The definition of composition in CPHTop and the proof of the factorization theorem essentiallyused the assumption that the index sets A be cofinite. On the other hand, the construction of the homology groups HP(X; G) did not require thisassumption. Therefore, it remained unclear whether one can use also non-cofmite ANR-resolutions to determine the homology groups Hp{X; G) of the space X. To prove that thisis indeed the case is the main purpose of this paper. Such an information can prove useful in situations where a non-cofmite ANR-resolution naturally arises. The main idea of the proof is to replace a given ANR-resolution p: X―>X by a cofinite ANR-resolution p*: X-*X* using the "trick" described in ([9], Theorem I, 1.2). What remains to be done is to exhibit a natural isomorphism m* : HP(X; G)-^HP(X*; G). The correct formula for u* is easily found. However, the formula for the inverse v* of u* is less obvious. Even more complicated is the verificationof the two equalities u^v^―1, v*u*=l. In order to simplify notations throughout the paper we omit the coefficient groups G, although all results hold for an arbitrary G.
- Published
- 1987