Such as his comment that started an entire outrage was that openbsd contained non-free software, then his opening argument about non-free firmware blobs, and then onto the ports tree urls.
Edit: from Theo de Raadt http://article.gmane.org/gmane.os.openbsd.misc/134986
He really does. I used to be involved in Emacs development back in the day and he frequently changed his opinion on things if you presented him with with a well-reasoned argument. Sure, he's a stickler for his principles, but he's not afraid to admit when he's wrong on practical details or unable change his mind about stuff if you give him a good reason to.
Seems like a good trait to have.
"If you think Stallman's information theshold is too low, maybe you need some more information ;)" - Love, the smiley face. Yes, in this case I think the e-mails prove his information threshold was too low and it shows. I think you need to read the whole exchange :)