Here is the text of Exercise:

2 a) Let $X$ be an ordered set. Show that the set of intervals

$\left[x, \rightarrow\right[$ (resp. $\left]\leftarrow, x\right]$)

is a base of topology on $X$; this topology is called the right (resp. left) topology of $X$. In the right topology, any intersection of open sets is an open set, and the closure of $\{x\}$ is the interval $\left]\leftarrow, x\right] $.

The above one was from English edition. I translated French edition and found the same text.

Should not be $X$ a totally ordered set ? And is not that the set of intervals should be $\left]x, \rightarrow\right[$ in place of $\left[x, \rightarrow\right[$ ?

Is this an errata ?

  • 3
    $\begingroup$ I looked it up, and "order" does, indeed, mean partial order. $\endgroup$ – Gerald Edgar Jan 30 '12 at 14:13
  • 5
    $\begingroup$ errata = a list of errors in a published work. So it's not an "errata", but possibly (at most) an error. But I don't think it is, it's correct as stated. $\endgroup$ – Henno Brandsma Jan 30 '12 at 14:27
  • 3
    $\begingroup$ erratum is the singular form. :) $\endgroup$ – Jim Conant Jan 30 '12 at 15:55
  • 2
    $\begingroup$ Still I would also suggest to avoid using erratum in this form, as it is my impression its standard meaning in academic writing is a bit different from the pure translation 'error'; often referring to some (informally) published note pointing out and possibly fixing an error. $\endgroup$ – user9072 Jan 30 '12 at 17:50
  • 5
    $\begingroup$ Somewhere there is a book in which page 342 is headlined "Errata", and the text reads, in its entirety: "On page 342, change 'Errata' to 'Erratum' ". $\endgroup$ – Steven Landsburg Jun 25 '13 at 4:02

Bourbaki was right :-)   On the other hand, let   $(X\ \le)$   be a partially ordered set. In general the family

$$\mathbf B\ \ :=\ \ \{\ ]x,\rightarrow[\ :\ x\in X\ \}$$

is NOT a topological base for any topology in   $X$.   One reason is trivial: no minimal element belongs to any member of   $B$;   thus if there is any minimal element then   $X$   would not be open.

OK, one could define:

$$\mathbf B\ \ :=\ \ \{X\}\ \cup\ \{\ ]x,\rightarrow[\ :\ x\in X\ \}$$

It will not help. Indeed, here is a characterization of a topological base:

THEOREM   A family   $\mathbf B$   of subsets of   $X$   is a topological base for a topology in   $X\quad\Leftrightarrow$   the following two conditions hold:

  •   $\bigcup \mathbf B\ =\ X$

  •   $\forall_{G\ H\in\mathbf B}\quad G\cap H\ =\ \bigcup\ \{K\in \mathbf B : K\subseteq G\cap H\} $

Now consider a 5-element set

$$X := \{b\ \ d\ \ A\ \ C\ \ E\}$$

where by definition there are exactly four sharp inequalities   $b < A$   &   $b < C$   &   $d < C$   &   $d < E$.   Then the intersection

$$ ]b,\rightarrow[\ \ \cap\ \ ]d,\rightarrow[\quad=\quad \{ C \} $$

is not a union of any family of open rays   $ ]x,\rightarrow[ $.

| cite | improve this answer | |
  • $\begingroup$ May I know, please, what's wrong with my answer? @Reetesh Mukul wondered if Bourbaki should consider open intervals instead of closed intervals as a base of the left (or right) topology. I tried to explain that one would not get a topological base for general partially ordered space. Is it a crime on MO? $\endgroup$ – Włodzimierz Holsztyński May 24 '13 at 0:47
  • $\begingroup$ I am asking because my "Answer" has received a down-vote. $\endgroup$ – Włodzimierz Holsztyński May 24 '13 at 1:39

Say we have a partially ordered set. What so you doubt? (1) The set of intervals $\left[x,\rightarrow\right[$ is a base for a topology. (2) Any intersection of open sets is open. (3) The closure of $\{x\}$ is $\left]\leftarrow,x\right]$. They all look OK to me...

| cite | improve this answer | |
  • $\begingroup$ I solved this Exercise using the idea of Poset only, but then I got confused after reading the definition of right and left topology in Wiki: en.wikipedia.org/wiki/…. In General Topology -I, 1.4, while giving definition of Closed Set, Bourbaki says: On the rational line, every interval of the form [a,→[ is a closed set, for its complement ]←,a[ is open. But I agree that closed/open sets are subject to Topology under consideration. $\endgroup$ – nature1729 Jan 30 '12 at 14:45

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.