Bem himself has said that he chose his hypotheses in advance. For you to state otherwise is to implicitly accuse him of deception.
You say that Bem said that he chose his hypotheses in advance. Can you present a quote where he said that, because
according to Linda, "Bem stated 'in fact, there was no data exploration
that required adjustment for multiple analyses in this or any other experiment' in response to Wagenmakers' question about whether the data was exploratory. [emphasis added by Jay]" That's not the same as saying that he chose his hypotheses in advance, especially when you take into account Bem's opinion on correction for multiplicity. As he stated in his textbook chapter, "Spurious findings can emerge by chance, and we need to be cautious about anything we discover in this way.
In limited cases, there are statistical techniques that correct for this danger [emphasis added by Jay]."
So he seems to believe (and to an extent he is correct) that usually there is no good statistical technique to correct for multiple analyses. Now go back to the previous paragraph and reread the claim that Bem actually made.
Furthermore, as I tried in vain to explain to Chris, whether the hypotheses were stated "in advance" is not the issue. Unadjusted p-values from testing multiple hypotheses that were stated in advance usually are not interpretable, and should either be adjusted formally or taken with a grain of salt.
And finally, I don't think that adjusting Bem's p-values is even the issue, because, for one thing, I agree with him, that if he did the sort of significance seeking that he appears to have done, that there is no good way to adjust the p-values (except, perhaps making them all p=1.0, since that's what the real significance level is). Rather, I think researchers need to avoid significance-seeking analyses in the first place.