Careful incremental steps are the way to proceed. Let me explain the current step I'm taking. People generally visit TakeOnIt from a Google search to find out opinions about a particular issue that they searched for. Let's say that they arrive at this page, to find out opinions on whether evolution is true:

http://www.takeonit.com/question/27.aspx

I've annotated some of these quotes - it's just a start - with "woos", or whatever we want to call them. Now let's say none of these quotes were annotated. The result? A person can be persuaded without seeing the general patterns of how they got persuaded. This happened to a friend of mine. He got persuaded by a quote where I was like: "you can't see why this argument is duping you?!". It took me a while to explain the persuasion tactics used in the quote. It was basically a slow process of me identifying and communicating persuasion patterns.

That's where labeling the quotes comes in. It allows a smaller community who's familiar with these patterns of persuasion to pre-process those patterns for the larger community. Now, there's an additional step we could take. We could categorize the various kinds of persuasion patterns. So X is an "alleged fallacy" while Y is "rhetoric" and so on. I actually suspect that more than one category per label is required and it's a mistake to think these labels naturally fall into discrete mutually exclusive categories. However, this label categorization step is secondary. The highest order bit is to simply label the persuasion in the first place. That's where most of the cognitive work is.

I considered allowing a truly broader notion than a "woo", which was simply to allow any tag at all on a quote. However, I think restricting the tags to persuasive patterns gives good focus and avoids dilution.