2

I'm trying to figure out how to prove or disprove the following statement:

Infinite union of recursive languages is recursively enumerable.

I know how to prove that infinite union of regular languages is not recursive, but I can't figure out how to prove the statement above. Intuition is telling me that the statement is true, but I'm not sure. Does anybody know how would I go about it?

Any help is greatly appreciated.

Raphael
  • 72,336
  • 29
  • 179
  • 389
flashburn
  • 1,223
  • 1
  • 12
  • 22
  • 3
    Consider the infinite usion of languages of the form ${w}$, i.e., each containing a single word. – Hendrik Jan Nov 12 '13 at 00:29
  • @HendrikJan: I think this will just tell me that infinite set of recursive languages is not recursive. What I need to show that infinite set is recursively enumerable (or not). – flashburn Nov 12 '13 at 00:32
  • 3
    Why would $\bigcup_{w\in L} {w}$ be RE? – Hendrik Jan Nov 12 '13 at 00:36
  • @HendrikJan: I realize that my answer might sound stupid, but would you mind explaining why it is not RE? My professor is not very good at explaining the material and I would greatly appreciate any help. – flashburn Nov 12 '13 at 00:38
  • Not every language $L$ is RE. – Hendrik Jan Nov 12 '13 at 00:41
  • @HendrikJan: Then my next question would be why infinite union of regular languages is not recursively enumerable? – flashburn Nov 12 '13 at 00:45
  • @HendrikJan Can you explain why I can make an assumption that a union of singletons is a non RE? – flashburn Nov 12 '13 at 02:16
  • @flashburn Not all infinite unions are created equal. There are countable unions, uncountable unions, enumerable unions, and probably more. – Raphael Nov 12 '13 at 16:51

1 Answers1

5

Choose an arbitrary language $L$ that is not RE. Denote $L_w=\{w\}$ the language containing a single word $w$. Clearly, $L=\bigcup_{w\in L} L_w$ is a union of infinitely many regular and recursive languages.

avakar
  • 626
  • 3
  • 12
  • I think I'm missing something. I realize that this question might sound strange to someone but as I said my professor is not very good at explaining the material. My question is why I can make an assumption that I can choose a non RE language such that it is a union of regular languages? – flashburn Nov 12 '13 at 13:39
  • @flashburn, I'm disproving the statement "for all infinite sets of recursive languages, their union is RE" by showing a counterexample. In this case, I'm saying "look, here's one concrete language $L$ and and infinite set of regular languages $L_w$ whose union is $L$". Sure, I didn't actually say what $L$ is, but that's not really a problem -- all you need to prove is that there is at least one (and we kinda know that there is :) ). – avakar Nov 12 '13 at 14:27
  • @flashburn This works for the following reason: (1) Not every language is RE; (2) Every language can be expressed as an infinite union of languages containing a single word; (3) therefore, there are languages that are not RE that can be expressed as an infinite union of languages containing a single word. Similarly: (1) Not all mammals are people; (2) All mammals are living things; (3) therefore, there are mammals that aren't people that are living things. – Patrick87 Nov 12 '13 at 20:18