James Dugundji defines (Topology, chap. VI definition 5.1) a cone in the following manner
For any space $X$, the cone $TX$ over $X$ is the quotient space $(X\times I)/R$, where $R$ is the equivalence relation $(x,1)\sim(x',1)$ for all $x$, $x'\in X$.
My question is on the relation $R$. How is $$R=\{\big((x,1),(x',1)\big)\mid x,x'\in X\}\subset (X\times I)^2$$ an equivalence relation on $X\times I$? It is not necessarily reflexive [$(x,0)\nsim(x,0)$]. What does Dugundji mean with that definition of $R$?
He then, in the next line,writes
Equivalently, $TX=(X\times I)/(X\times 1)$; intuitively, $TX$ is obtained from $X\times I$ by pinching $X\times 1$ to a single point.
And I, again, don't know what he means: how is he taking the quotient space of $X\times I$ with respect to $X\times 1$ which is not a subset of $(X\times I)^2$?
Notation:
- Dugundji uses $x=\{x\}$ sometimes (in this case, $x=1$).
- $I=[0,1]$.
I guess he means the equivalence relation generated by $(x,1)\sim(x',1)$, i.e. the smallest equivalence relation containing all couples $((x,1),(x',1))$. In this case, that means that $(x,a)R(y,b)$ only if $(x,a)=(y,b)$ or $a=b=1$.
In the next line, he considers $X\times\left\{1\right\}$ as a subset of $X\times I$ and he takes the quotient topology of this subset. That is, he lets two points in $X\times I$ be equivalent if and only if they are either equal or both in $X\times\left\{1\right\}$.