2

Any answers will be appreciated. I know $\mathbf Q$ has closure $\mathbf R$, but can $\mathbf Q$ be expressed as the sum of two nowhere dense sets? Or will some other example work?

user153012
  • 12,890
  • 5
  • 42
  • 115
adrija
  • 3,097
  • You can make $A$ to be the integers and $B= \sqrt{2} A$ – PenasRaul Sep 01 '14 at 21:58
  • @PenasRaul This does not seem to work because $A+B\neq\mathbb{Q}$ – user2097 Sep 01 '14 at 22:01
  • @user2097 An additive subgroup of $\mathbb{R}$ is either cyclic or dense. Since both $\mathbb{Z}$ and $\sqrt{2}\mathbb{Z}$ are additive subgroups and their sum is not cyclic, their sum is dense. On the other hand both are closed with empty interior. – egreg Sep 01 '14 at 23:25
  • @egreg The OP wants $A+B=\mathbb{Q}$ if I understand the question correctly. – user2097 Sep 01 '14 at 23:29
  • @user2097 My impression is that the title and the question don't match and that asking about $\mathbb{Q}$ is just an attempt. And there is “Or will some other example work?” – egreg Sep 01 '14 at 23:33
  • @egreg You are right. Now I see that OP asked two different questions. – user2097 Sep 01 '14 at 23:43

2 Answers2

6

$A+B=\mathbb{Q}$ holds with $A=\mathbb{Z}$ and $B=\{q_n-n\}$, where $q_1,q_2,\ldots$ is an enumeration of $\mathbb{Q}\cap[0,1)$.

user2097
  • 2,455
0

A well-known property of the Cantor set (which is nowhere dense) is that $C+C=[0,2]$. (See here.)

Now, let $K$ be the 1-periodic continuation of Cantor set, i.e., $$ K=\bigcup_{k\in\mathbb Z}k+C. $$ Then $K$ is nowhere dense, and $$ K+K=\mathbb R. $$