In general, we know that adding a subset to a regular cardinal $\kappa$ can collapse cardinals. If, for example, there is $\gamma < \kappa$ with $2^\gamma >\kappa$, then $Add(\kappa,1)$ will collapse $2^\gamma$ to $\kappa$, since every subset of $\gamma$ will appear as a block in the generic.

However, this argument requires that we use the poset defined in $V$. We can easily construct situations in which $2^\gamma > \kappa$ but forcing with $Add(\kappa,1)^L$ does not collapse cardinals, by the same techniques that allow us to achieve Easton's Theorem. If we force over $L$ with $Add(\gamma, \kappa^+) \times Add(\kappa,1)$, for example, then cardinals are preserved. I am interested to know if this holds in general.

Does $Add(\kappa,1)^L$ ever collapse cardinals?

  • $\begingroup$ What is the partial order for $Add(\kappa, 1)$? $\endgroup$ May 22 '15 at 4:25
  • $\begingroup$ @Nate: It's the forcing $\operatorname{Add}(\kappa,1)$ (all the partial functions from $\kappa\to2$ with domain bounded below $\kappa$; ordered by reverse inclusion), but only those functions which are in $L$. $\endgroup$
    – Asaf Karagila
    May 22 '15 at 6:01
  • 1
    $\begingroup$ (I removed the [cohen-forcing] tag, I think it's overly specific.) $\endgroup$
    – Asaf Karagila
    May 22 '15 at 6:02
  • $\begingroup$ @Asaf: Thanks. But then this raises three other questions. First, why is this $Add(\kappa, 1)$ and not $Add(\kappa, 2)$? What is $Add(\gamma, \kappa^+)$? And then what is the difference between $Add(\kappa, 1)$ and $Add(\kappa, 1)^L$ (which is mentioned in the question as separate)? $\endgroup$ May 22 '15 at 7:55
  • $\begingroup$ @NateAckerman Adding one Cohen subset to $\kappa$ is isomorphic to adding two such sets, so $\text{Add}(\kappa,1)\cong\text{Add}(\kappa,2)$. In general, $\text{Add}(\kappa,I)$ should consist of partial functions on $\kappa\times I\to 2$, with support of size $\kappa$. In the case of $1$, we simply suppress the extra coordinate as unnecessary. The superscript $L$ on the notation $\text{Add}(\kappa,1)^L$ means that Jonas is considering the version of the forcing notion as defined in $L$ (by taking only such conditions that are in $L$), but he wants to force with that poset overe $V$. $\endgroup$ May 22 '15 at 10:52

Yes, it follows from a theorem of Mack Stanley (for this special case, in fact from a theorem of Foreman-Magidor-Shelah) that if $0^\sharp$ exists, then forcing with $Add(\kappa, 1)^L$ collapses $\kappa$ into $\omega.$

See Stanley's paper "Forcing disabled" (and Foreman-Magidor-Shelah paper "$0^\sharp$ and some forcing principles", J Symbolic Logic 51 (1986), 39-46). See also Can the Cohen forcing collapse cardinals? and the answer given there.


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.