You know, explaining an idea to somebody, from scratch, it's the better way to develop it. If it's stupid they will realize. If it's good they'll get enthusiastic.
In a previous comment I mentioned that this is the tactic that my manager uses, and it works wonderfully: let people talk from the beginning and ask questions later.
Of course, if you end up disagreeing, even though if their idea is good, they'll like that you let them explain and took time to consider
Part of the use case of a "trial by fire" approach is that the idea is to come up with a better approach than any one of us could ideally. One isn't interested in trying to use their idea as-is, nor in simply disregarding it, but in probing, understanding, and adapting it to the case at hand.
That means arguing about it in detail. Explaining from scratch may be a necessary prelude but it doesn't help really with the final deliberative process. The key is to ensure that everyone knows that these are good-natured and that most people will shape the outcome even if their ideas are not adopted wholesale.
The goal is to have an argument where everyone collectively wins, and where each reasonable participant (meaning, most of the time, every individual participant) individually wins as well, in the sense of having their concerns incorporated into the final view.
1) Those who know what's going on and treat it as a sign of respect.
2) Those who don't and where you have to show externally that there is respect involved.
For the first group, getting into an argument on technical merits doesn't have to involve a whole lot of pleasantries. It's largely informal. I try to save this for groups I am a long-standing member of.
For the second group, you are right, a softer approach is necessary.
However, you can also wrap that in a friendly approach as youlost_thegame mentions - simply acknowledge the potential of the idea first (ie, positive energy) then go in with the tough questions.
I find even with friends and close colleagues, if I stick to the "nice first, then difficult questions combined with enthusiasm", I can often critique the idea while making the entire conversation positive. Often the faulty idea gets discarded by it's owner after consideration.