It seems to me that arguments between scientists are productive mostly because they have a lot of shared context. If the goal of arguing is to learn things for yourself, then it's useless to argue with someone who doesn't have the relevant context (they can't teach you anything) and useless to argue about a topic where you don't know the relevant context yourself (it's better to study the context first). Arguments between people who are coming from different contexts also seem to generate more heat and less light, so it might be better to avoid those.