Topological Subspaces

Just like you can define a sub-set of a set, or a sub-object of an object in a category, you can define a sub-*space* of a topological space. It's a pretty easy thing to understand; interestingly, a sub-space of a topological space works in pretty much exactly the same way as a sub-sets and sub-object. In fact, the topological definition of a sub-space is *identical* to the categorical definition of a sub-object when we're looking at the category of topologies, **Top**.

Today, I'm going to explain what a subspace is, and show you how the categorical sub-object corresponds to the topological subspace. Read on beneath the fold.

Let's start with the normal topological definition. Suppose we have two topological spaces, (T, τ), and (S,σ). (S,σ) is a sub-space of (T,τ) if and only if:

1. S ⊂ T
2. For each open set Os ∈ (S,σ), ∃ open set OT ∈ (T,τ): Os = OT ∩ S.

So saying (S,σ) is a subspace of (T,τ) means that S has a subset of the objects that are in T, and (S, σ) structures the objects that it contains in the same way as T; so (S,σ) preserves as much of the structure of (T,τ) as you can represent using the objects contained in S.

Using that definition, given a set of objects, S ⊂ T, we can construct a topological space *induced by* T by generating open sets using statement 2 from the definition of subspaces. We call the set of open sets σ the *relative topology* of T on S. The neighborhoods in the relative topological space (S,σ) are called the *relative neighborhoods* of S.

Another way of defining the relative neighborhoods without generating the induced space is to apply statement 2 from the definition of subspaces in the definition of relative neighborhoods: Suppose we have a subspace (S,σ) of a topological space (T,τ); and let o ∈ S. Then *N(o)* is a relative neighborhood of *a* if/f: ∃ a neighborhood *M(o)* in (T,τ) such that *N(o)= M(o) ∩ S*.

We can also define *relatively closed* sets; it's exactly the same trick of pulling statement two of the definition of subspaces into the definition of closed sets. Given a subspace (S,σ) of (T,τ), the *relatively closed* subsets of S are the sets C ⊂ S such that for *some* closed set D in (T,τ), C = D ∩ S.

For all subspaces (S,σ) of the topological space (T,τ), a function f : T → X is continuous if and only if f restricted to S is continuous on (S,σ). (f restricted to S is a function g : S → X;: ∀ s ∈ S, g(s) = f(s).).

---------

Since we'll use some category theoretical stuff in some later topology posts, let's take a look at how this corresponds to the categorical concept of sub-objects. The set of topological spaces forms a category **Top**, where the *objects* in Top are topological spaces; and the *arrows* between them are continuous functions.

To review a bit, in category theory, a sub-object is defined in terms of *monic arrows*. A monic arrow is the category theoretic version of an injective function: f: a → b is *monic* if/f for all other arrows g and h : x → a, fº g = f º h implies that g = h. In other words, a monic arrow will *only* map two other arrows to the same place if those two other arrows are the same.

Using monic arrows, we can define an *equivalence class* of arrows. Suppose we have two monic arrows f : b → a, and g : c → a. If there is an arrow h such that g º h = f, then f ≤ g. If f ≤ g and g ≤ f then f ≡ g.

Each equivalence class defines a set of sub-objects of a; and those sub-objects are treated as the same object, because with respect to a, they are indistinguishable.

So - let's think about this in the category of topological spaces. A *monic* arrow f : s → t, where s = (S,σ) and t = (T,τ) in the category of topological spaces is an *injective* function - that is, it maps every object in S to a distinct object in T. That is, every object in S can be identified with exactly one object in T. Further, there is a preservation of structure: all arrows that can be mapped through the sub-object s arrive at the same mapping in T; that is, if we have a continuous function from X to T, and we use it on S, it will always generate the same result as if we applied it to *T* directly. Since for the function to still be continuous on S (which it must be to be an arrow), that means that for all possible continuous functions that can be applied to S, S will behave exactly the same as T; preserving structure the same way that T does. So the open-set relationships will be preserved exactly as in the topological definition above.

Tags

More like this

With continuity under our belts (albeit with some bumps along the way), we can look at something that many people consider *the* central concept of topology: homeomorphisms.
Back when GM/BM first moved to ScienceBlogs, we were in the middle of a poll about the next goodmath topic for me to write about. At the time, the vote was narrowly in favor of topology, with graph theory as a very close second.
Yesterday, I introduced the idea of a *metric space*, and then used it to define *open* and *closed* sets in the space. (And of course, being a bozo, I managed to include a typo that made the definition of open sets equivalent to the definition of closed sets.
I'm going to start moving the topology posts in the direction of algebraic topology, which is the part of topology that I'm most interested in.

Your definition of subspace should probably include something like "for every open set U in T, U \cap S is open in S"--otherwise, a space with the trivial topology would be a subspace of practically everything!

Subspaces do NOT correspond to categorical subobjects in TOP. Otherwise, |R trivially topologized (i.e. points are open) would
be a subspace of |R with the usual topology.