[kunglao]

The problem with trying to be "define any assumptions" and using "a clear chain of logic" is that, if you do this right, you will realise that there's really very few things you can actually say about anything at all. This is a problem when it comes to day to day dealings.

On the one hand, average population do not agree on what can be construed as "self-evident". On the other hand, what seems self-evident within our limited capacity is often misguided or frustratingly incomplete. Even if we sorted both these out, being logical means having to deal with details. My experience has been that average person does not like to deal with the details.

[tabtab]

First you find out what each side agrees on and what they don't. Part of the "logic" exercise is teasing out such info. Both sides may not agree in the end, but at least have a better idea of what assumptions (givens) each is using, and maybe test those once narrowed.

[toasterlovin]

> you will realise that there's really very few things you can actually say about anything at all

This really only applies if you require absolute certainty. But most of the time being reasonably certain is enough to operate in the world. And you can be reasonably certain about lots of stuff.

[tabtab]

Ask each side to give a probability. If both sides have a similar estimate, then you move on to the next logic step(s). Don't bother debating things you already agree on. We don't have to dissect everything.

If it varies too much, then explore the reasons for the different estimates. Often times it comes down to different past experience, which is obviously going to vary per individual. (Ex: "when users see X on UI screens, they usually do Y, in my experience".) But at least one can narrow down recommended decision logic to observed patterns of history. Narrowing down the point of disagreement is often still helpful even if both sides still disagree.