According to my lecture notes, with the help of a little "trick" it is possible to write an infinte union of sets as the infinite union of disjoint sets as follows:

$$A = \bigcup \limits_{k=1}^{\infty} A_k = \bigcup \limits_{k=1}^{\infty} \tilde{A_k}$$

with the disjoint sets $\tilde{A_k}$ whereas $\tilde{A_1} := A_1$ and $\tilde{A_k} := A_k \setminus A_{k-1}$ for $k \geq 2$.

My question: Given we have three sets $A_1,A_2$ and $A_3$ and none of them is disjoint. When I define $\tilde{A_3}$ as $A_3 \setminus A_2$ I don't see how $A_1$ and $A_3$ are supposed to be disjoint..so on what am I missing out?

  • $\begingroup$ It seems like it should be $\tilde{A}_k = A_k \setminus \tilde{A}_{k-1}$, with a tilde on the right hand side. $\endgroup$ – Dan Shved Mar 6 '13 at 18:37
  • $\begingroup$ Hm. Doesn't seem to be correct neither. $\endgroup$ – TestGuest Mar 6 '13 at 18:44
  • $\begingroup$ you are right, I haven't thought this through. What I should have said is in coffeemath's answer. $\endgroup$ – Dan Shved Mar 6 '13 at 18:47
  • 1
    $\begingroup$ It could also be that $A_i$ in your lecture notes are an increasing chain, i.e. $A_i \subseteq A_{i+1}$. If so, the definition makes perfect sense as it is. $\endgroup$ – Dan Shved Mar 6 '13 at 18:48
  • $\begingroup$ !! Ah, it is true, it is a chain. Sorry then, the property of the sets bein ordered as chain was half a page above! $\endgroup$ – TestGuest Mar 6 '13 at 18:49

The definition of $\bar{A^k}$ should be $A_k \setminus (A_1 \cup ... \cup A_{k-1}).$

One way to think of $\bar{A^k}$ is that it consists of all points (if any) which first appear in set $A_k$, not yet having appeared in any of the previous sets $A_1,...,A_{k-1}.$ To me this makes it clear the $\bar{A^k}$ are pairwise disjoint, with the same union as the original collection of $A_k$ sets.

  • $\begingroup$ Yeah, I do understand the intuitive idea..but still am not sure about the definition. So according to you the definition is entirely wrong? That would be a bit odd since it is used throughout the whole lecture notes. $\endgroup$ – TestGuest Mar 6 '13 at 18:46
  • 3
    $\begingroup$ The definition in the question would have been correct if we were dealing with an increasing union, i.e., if $A_k\subseteq A_{k+1}$ for all $k$. Perhaps the author had this assumption in mind and neglected to state it. $\endgroup$ – Andreas Blass Mar 6 '13 at 18:49
  • 1
    $\begingroup$ It was stated, it was my fault that I missed out on it. Sorry! $\endgroup$ – TestGuest Mar 6 '13 at 18:54

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.