# Understanding Hatcher's definition of CW-complex

In Hatcher's book on Algebraic Topology, the author writes:

(1) Start with a discrete set $X^0$ , whose points are regarded as $0$ cells.

(2) Inductively, form the $n$ skeleton $X^n$ from $X^{n−1}$ by attaching $n$ cells $e_{\alpha}^n$ via maps $\phi_{\alpha}:S^{n-1} \to X^{n-1}$. This means that $X^n$ is the quotient space of the disjoint union $X^{n−1} \cup \cup_{\alpha} D_{\alpha}^n$ of $X^{n−1}$ with a collection of $n$ disks $D_{\alpha}^n$ under the identifications $x ∼ \phi_{\alpha}(x)$ for $x \in \partial (D_{\alpha}^n)$ (...)

A space $X$ constructed in this way is called a cell complex or CW complex.

I am having many problems understanding this definition...

1) What is a $n$-cell? (the author does not define it previously)

2) What does he mean by "attaching $n$ cells $e_{\alpha}^n$ via maps $\phi_{\alpha}:S^{n-1} \to X^{n-1}$"? What is the relation between the $e_{\alpha}^n$ and $\phi_{\alpha}$?

3) How can one realize the given definition with the description - "This means that $X^n$ is the quotient space of the disjoint union $X^{n−1} \cup \cup_{\alpha} D_{\alpha}^n$ of $X^{n−1}$ with a collection of $n$ disks $D_{\alpha}^n$ under the identifications $x ∼ \phi_{\alpha}(x)$ for $x \in \partial (D_{\alpha}^n)$ (...)"

I know this is an important concept to grasp in Algebraic Topology, so I would really like to understand it, but this definition wasn't clear at all to me.

I would post it as a comment, but I really want to draw a commutative diagram... Categorically speaking, $X^n$ is a pushout in the category of topological spaces

$$\require{AMScd} \begin{CD} \coprod_\alpha S^{n-1}_\alpha @>{\coprod_\alpha i_\alpha}>> \coprod_\alpha D^n_\alpha \\ @VV (\phi_\alpha) V @VV V\\ X^{n-1} @>>> X^n \end{CD}$$

Here the arrow $\coprod_\alpha i_\alpha$ is formed by the standard inclusions of $(n-1)$-spheres as boundaries of $n$-disks, and $\phi_\alpha\colon S^{n-1}_\alpha \to X^{n-1}$ are some maps that you choose. The description of $X^n$ as a disjoint union of $X^{n-1}$ and $\coprod_\alpha D^n_\alpha$ with certain identifications is the description of pushouts for topological spaces.

Then an $n$-cell is each map $D^n_\alpha \to X^n$ given by the pushout above.

1) A n-cell is a map from $$D^n$$ to the topological space $$X$$ you're working with.

2) Attaching a space $$X$$ to a space $$Y$$ along a map $$\phi : E \to Y$$, assuming $$E \subseteq X$$, means to create the quotient space over the disjoint union of $$X$$ and $$Y$$ by identifying a point $$e \in E$$ with its image $$\phi(e)\in Y$$. Figuratively, we "stitch" $$X$$ upon $$Y$$ by gluing $$E \subseteq X$$ onto $$Y$$ with $$\phi$$ and we obtain the space

$$\frac{X \coprod Y}{ \langle e \sim \phi(e) \rangle}.$$

So far you've constructed you cellular complex up to $$n-1$$-cells, obtaining the space $$X^{n-1}$$. View this as the "skeleton" upon which you will attach a $$n$$-disk $$D^n$$ along the boundary of the disk $$\partial D^n = S^{n-1}$$. Thus you need that map $$\phi^n_\alpha : S^{n-1} \to X^{n-1}$$ to identify the boundary of the disk as points of the skeleton $$X^{n-1}$$.

$$e^n_\alpha$$ is the cell, i.e. the map that send the entirety of the disk $$D^n$$ to the completed space $$X$$ (or possibly $$X^n$$), but not to $$X^{n-1}$$ as the point in the interior of the $$n$$-disks are not identified with any point of $$X^{n-1}$$. However on the boundary of the disk, $$e^n_\alpha$$ should match $$\phi^n_\alpha$$ :

$$e^n_\alpha|_{\partial D^n} = \phi^n_{\alpha}.$$

3) I already explained the concept of attaching spaces in 2). From that definition,

$$X^n = \frac{ X^{n−1} \cup \big( \bigcup_{\alpha} D_{\alpha}^n \big) }{\langle x_\alpha \sim \phi^n_\alpha(x_\alpha) \rangle}$$ holds, with the relation being spanned over every $$x_\alpha \in S^{n-1}_\alpha$$ for any index $$\alpha$$.

Note that Hatcher uses another notation than the one I used earlier to introduce the concept, where rather than using the disjoin union $$\coprod$$, he uses the regular union $$\bigcup$$ and label each of its disks to be attached with a unique index $$\alpha$$ so they are still distinct after the union. The topological result is the same.

Regarding the OP's question 1, while the accepted answer gives a valid interpretation of an $$n$$-cell in a CW complex, it is worthwhile to note that this interpretation contradicts the meaning of $$n$$-cells stated in Hatcher's textbook. Namely, in Hatcher's book, the $$n$$-cells $$e^n_\alpha$$ are not maps to the CW complex defined on the closed $$n$$-disc, but instead subsets of the CW complex homeomorphic to the open $$n$$-disc.

First, the OP's quotation from Hatcher contains an ellipsis "(...)" which elides an important portion of Hatcher's definition in Chapter $$0$$:

Thus, as a set, $$X^n = X^{n-1} \coprod_\alpha e^n_\alpha$$ where each $$e^n_\alpha$$ is an open $$n$$-disk.

Second, shortly after this passage Hatcher writes

The reader who wonders about various point-set topological questions lurking in the background of the following discussion should consult the Appendix for details.

and then, upon consulting the Appendix, one find a variation on the "(...)" sentence:

The cell $$e^n_\alpha$$ is the homeomorphic image of $$D^n_\alpha - \partial D^n_\alpha$$ under the quotient map.

Finally, the maps defined on closed $$n$$-cells do also occur in Hatcher's later discussion in Chapter 0, under the terminology of "characteristic maps":

Each cell $$e^n_\alpha$$ in a cell complex $$X$$ has a characteristic map $$\Phi_\alpha : D^n_\alpha \to X$$ which extends the attaching map $$\varphi_\alpha$$ and is a homeomorphism from the interior of $$D^n_\alpha$$ onto $$e^n_\alpha$$.