I wrote before about shoulding and the ways that obligatory thinking is pathological. Locked inside every should is either something that you want to be doing, or something that you don't actually have a good reason to do. Unfortunately, focusing on the obligation makes it hard to see whether there is any motivation of substance underneath. Worse still, I've realised lately that shoulds have a few ways of propagating into other areas and obscuring your motivations there too.
The first is that shoulds propagate through dependencies. If you feel obligated to take out the garbage, but first you are going to finish reading your book, the obligation spreads: now you have to finish your book so you can take out the garbage. Maybe you truly enjoy the book, but if you don't really want to take out the garbage then every page you read is bringing you closer to an unpleasant situation. On the other hand, there's also now a degree of pressure on finishing the book that you didn't have before. Reading it used to be something you wanted to do, now it's something you have to do.
The second way is through substitution. It's not uncommon that you will bump one activity for another, better activity. With an obligation-free activity, that's no problem. I wanted to read, now I want to go for a walk. But when a should is involved, it propagates implicitly: you must be at least as obligated to do this new thing as the old thing it replaced! I understand you can't come to my party because you have to work; I recognise your superior obligation. But if you ditch my party because you're too tired, I'd better not find you online later in the night. You are obligated to sleep by substitution!
There are a couple of ways I can think of to mitigate these. I covered reducing dependencies before, but another possibility is to just reorder things as much as you can, so the shoulds happen first and can't contaminate anything else. I covered something like that in motivation-driven development. As for mitigating substitution, perhaps the best I can suggest is to separate the two decisions: drop the old thing unconditionally, then pick up the new thing.
But all this is just more reason to avoid shoulds, in my opinion.