- Start with a clearly-defined problem of the form “It looks like so-and-so is the case but it also looks like so-and-so is NOT the case.” So for example “It looks like democracy is good because it lets people have a voice in their own fates, but it also looks like democracy is not good because democracies often vote to do bad things, including electing dictators who end the democracy.” The writing will be better if this problem is clearly defined and real — that is if you yourself and the people reading it see a good reason for adopting both sides of the position.
- Resolve the problem by making a proposal that would cost something to adopt. That is make sure that what you say could be wrong, that assuming it to be true causes us to run some sort of clear risk. So for example, to follow our example of democracy you might propose that democracies are only good if they never infringe on the rights of minorities, and you go on to say in a clear way that a sane person could disagree with what counts as a minority and what counts as infringing on the rights of minorities. Not all risks are political, obviously. So if you are arguing the case that people should pay attention to poetry, and considering the counter-argument that it is a waste of time, you might conclude people should pay attention to poetry even if it is a waste of time. So you would be advocating running a risk of wasting time.
- Give your reader a chance to look at all the points AGAINST what you are saying. So in the case above be honest that you are limiting democracy, and also be honest that many societies that protect the rights of minorities have ended badly, and be honest about what the limits are on what sort of rights of minorities you would protect — e.g. would you protect the rights of Christian Scientists to deny medical care to their children?
- Do the best you can only to use words everybody knows and only to appeal to facts that everybody could easily check.