I was thinking about some topological ideas, especially those relating to shape theory, and came across some interesting constructions that seem to relate to shape theory, but that I can't quite situate within my own knowledge - so was wondering if there is any related ideas out there in math.

Let $\beta$ be the Stone-Cech compactification functor. For any space $S$, let us define a new space $CS$ as follows: Consider the space $\beta(S\times [0,1))$ and the projection $\beta \pi_2:\beta(S\times [0,1))\rightarrow [0,1]$ given by extending the projection to the second coordinate. Define $CS$ to be the fiber of $1$ under $\beta \pi_2$ - note that if $S$ is itself compact Hausdorff, this is just $CS=\beta(S\times [0,1))\setminus S\times [0,1)$. This space somehow "expands" $S$ to be large enough that any map $S\times [0,1)$, when thought of as a homotopy missing its endpoint, will in some sense have a limit.

More precisely, if $Q=\prod_J[-1,1]$ for some indexing set $J$ with projections $\pi_j:Q\rightarrow[-1,1]$ for each $j\in J$ and $B$ is some compact subset of $Q$, one may explicitly describe the maps $CA\rightarrow B$: they are continuous functions $f:S\times [0,1)\rightarrow Q$ such that for any open set $N$ containing $B$, there is some $\alpha$ such that if $t>\alpha$ then $f(a,t)\in N$ under the equivalence relation that $f\sim g$ if for every $j\in J$ and $\varepsilon>0$ there is some $\alpha$ such that if $t>\alpha$ then $\pi_j(f(a,t))-\pi_j(g(a,t))<\varepsilon$. Note that every compact Hausdorff space $B$ can be embedded in some such $Q$.

This characterization seems very similar to the definition of morphisms in Borsuk's Shape Category except that morphisms in that category are taken up to homotopies, whereas these continuous functions $CA\rightarrow B$ are more strictly continuous maps - I haven't worked out all the details, but I think that under an appropriate notion of equivalence (e.g. some sort of homotopy in the corona of $A\times [0,1)\times [0,1]$ using the two embedding of $CA$ into this with last coordinate $0$ and $1$), one can get "morphism" in the shape category as an equivalence class of maps $CA\rightarrow B$.

This leaves me with a natural question: can one use this construction to handle some of the nasty spaces (e.g. the Warsaw circle) that shape theory handles without working in a category defined only up to homotopy? More explicitly, suppose we have some set of reasonably nice spaces (maybe something like compact Hausdorff) - can we define a category whose morphisms from $A$ to $B$ are continuous maps $CA\rightarrow B$? The stumbling block is that there's no obvious composition law to use - but maybe there is some clever composition law, or perhaps $C$ can be given the structure of a comonad via some map $CA\rightarrow CCA$ (although no such maps come to mind!), or maybe the explicit description of the maps $CA\rightarrow B$ can somehow be used (although it seems to become difficult to show well-definedness when one tries to do this).

Is there a way to make a category of "shapes" not up to homotopy along these lines? Is there any literature related to this idea?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.