I agree, they are using two different definitions of "good listening". I'd say the "fit in with the social interaction" one is usually the most relevant. This reminds me of Wittgenstein's "language games"[1]. Maybe I'm butchering Wittgenstein's thought, but my understanding is that language works in a social situation as a game, as an activity where things "work ok" or don't work ok. It's not about me communicating my inner mental state and you making sure you're understanding my inner mental state (Wittgenstein's argument is that this is not generally possible, but also maybe it's not even what we usually care about)

Of course, when you're in a specific setting (such as doing science or a police investigation) the other definition can be more relevant.

[1] https://en.wikipedia.org/wiki/Language_game_(philosophy)#:~:....