MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

If we assume that $\Bbb R^V\neq\Bbb R^L$, can we deduce that there is some $x\in\Bbb R^V$ which is $L$-generic?

Of course if $V$ is a generic extension of $L$ this is true, but if $V=L[0^\#]$ this is also true (the existence of $0^\#$ implies the existence of $L$-Cohen generics), but $0^\#$ cannot be added by forcing. So it is possible to have reals which are not generic over $L$, but their existence does imply that a generic exist.

Is this always the case, or is there a counterexample? Namely, can we have $V=L[x]$ for some real number $x$ such that no real number in $L[x]$ is $L$-generic?

Edit: As Joel's answer shows, we can generate a counterexample using class forcing over $L$. To avoid these, we might as well require that the universe is not a class-generic extension of some $W$ for which $\Bbb R^W=\Bbb R^L$.

share|cite|improve this question
My knowledge of generics is pretty weak, but from the bit of recursion theory I know, and what I know of the analogies, it seems like some version of the usual means for constructing reals of intermediate degree ought to provide such an $x$? – Steven Stadnicki Aug 15 '13 at 20:58

Great question!

In order to formalize the notion, let us understand the phrase "$x$ is $L$-generic" to mean: there is some partial order $\mathbb{P}\in L$ and some filter $G\subset\mathbb{P}$ that is $L$-generic, such that $x\in L[G]$. In particular, this refers to set-sized forcing only.

In this case, we can give a negative answer. Sy Friedman has a way to undertake the coding-the-universe forcing in such a way that the generic extension $V[G]=L[R]$ is minimal: Minimal Coding, Annals of Pure and Applied Logic, 1989, pp. 233-297. In particular, every real in the extension is either in $V$, or generates $R$, which is not set-generic (although it is generic for class forcing). Thus, if you start in $L$ and then undertake this forcing, you get a class forcing extension in which there are no $L$-generic reals for set forcing, and indeed, no $L$-generic sets of any kind.

In the linked paper, Friedman states the following:

Corollary. There is an $L$-definable forcing for producing a real $R$ which is minimal over $L$ but not set-generic over $L$.

It follows that this extension $L[R]$ has no set-generic objects at all not in $L$. That is, it is a strong counterexample, which applies not just to reals, but to sets of any size.

share|cite|improve this answer
That sure answers the question, thanks! Now comes the obvious follow up, what if we allow class-generics? – Asaf Karagila Aug 15 '13 at 21:38
Well, that seems far more difficult. First, it cannot really be formalized in ZFC. Second, if you forbid class forcing, then our hands are too-much tied for building a counterexample. – Joel David Hamkins Aug 15 '13 at 21:42
Sy has also studied a concept of hyperclass forcing, but I don't know whether it provides a counterexample for the follow-up question. If it does, then there's an obvious further follow-up. – Andreas Blass Aug 15 '13 at 21:58
Joel, here's a slightly more sensitive follow up. Can we characterize (except trivially) models of the form $L[x]$ where $x$ is a real number, in which there are no $L$-generic reals? – Asaf Karagila Aug 16 '13 at 0:26
I seriously doubt there will be a good characterization of such models, since the coding-the-universe argument shows that any model $V$ is a ground model of such an $L[x]$. So our characterization will have to include in some a way an account of all possible $V$'s. – Joel David Hamkins Aug 16 '13 at 0:31

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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