I think the important thing isn't so much what you say as how you encourage third parties to spin it.
For example, IMO, the 10:23 campaign did a good job of branding itself. It came across as cheeky, cheerful disrespect (with a serious underlying message).
Sad though it may be [to accept this], gently taking the mickey out of someone is a vastly more effective strategy than engaging them in serious debate. Not because it makes the point any better, but because it is looked on more kindly by the framers in the media.
A rule-of-thumb used by many people is that, if you feel passionately about something, that is a point against your position. I can understand why they feel this way - in my experience, deeply-held beliefs are actually less likely to be right, because emotion tends to inhibit skepticism.
It's a pain in the arse for me, though, because I always aim to practice what I preach. So every debate I get into is (on some level) deeply personal. For example:
- If someone presents an argument for God that I can't refute, I'll spend one day a week in church for the rest of my life.
- If someone presents an argument for homeopathy that I can't refute, I'll use it as a treatment for my own conditions.
- If someone presents an argument against global warming that I can't refute, I'll... actually I have no idea what I'll do, but I'm sure it'll be significant.
This leads to a rather serious outlook on these matters.
As mentioned, though, I've come to the conclusion that seriousness is simply not an effective strategy if I actually want to convince people. It's a paradox: if you care, you must appear not to care.
As such, I'm going to make a change to my hierarchy of debate. Step #1 should now be split in two:
1) If someone has clearly done less reading up on the subject than me, I will try to judge whether they are willing to accept correction.
a) If so, I will gently try to explain what they've got wrong.
b) If not, I reserve the right to cheekily, cheerfully, ruthlessly take the mickey out of their dogma. Or just to ignore them completely.
Which approach I take will depend on whether there's anyone listening who might be influenced by the discussion, and how bored I am at the time.