"I always know what they need"

Part of the grievance is a comparison. You notice when they're down. You remember their mother's birthday. You can tell when work is getting to them, and you make the tea or take the children out. If you can do it for them, it seems fair to ask why they can't do it for you.

Some people are more attentive than others, and you may well be the more attentive one in your household. I don't doubt it. But it's worth testing how accurate the attentiveness is, because there's good evidence that people are worse at this than they believe, including with spouses.

Tal Eyal, Mary Steffel and Nicholas Epley ran twenty-five experiments on the question. They were testing the common advice to put yourself in the other person's shoes. People were asked to predict what another person was feeling or thinking, or what they'd prefer. Some were first told to take the other's perspective and imagine the world as they saw it. Some of the experiments involved spouses predicting each other's preferences.

Beforehand, a large majority of people believed that taking the other's perspective would make them more accurate. It didn't. If anything it made them slightly less accurate, while sometimes making them more confident. Only one thing in the whole series improved accuracy, and it was in the final experiment. People were allowed to ask the other person, in conversation, and then they got it right more often. The authors' summary is that understanding another mind requires getting their perspective, and that trying to take it doesn't work.

So when you believe you know what your partner needs, you're doing what they fail to do for you, which is inferring. You may be better at it than they are. You're very unlikely to be as good as you think. There may be things they've wanted from you for years that you've never picked up. You wouldn't know, because nothing tells you when you've missed one. Your partner may have a list of their own, of times you made the tea when they wanted to be left alone, or gave advice when they wanted sympathy.

There's an odd finding from research on gifts that shows how much we prize guessing over asking. Morgan Ward and Susan Broniarczyk studied people buying from a gift registry, which is a list the recipient has made of exactly what they'd like. Close friends were more likely than distant acquaintances to ignore the list and choose something else. They wanted the gift to show how well they knew the person. The recipients were left with presents they hadn't asked for, from the people who cared most. The distant acquaintances bought from the list and gave better-matched gifts.

Both of you are guessing, then, and both of you are overconfident. Each of you privately thinks you're doing the understanding for two. A couple can start asking each other without either of them having failed at anything.