I shall merely say that there are ways to divorce it from the context, at least in part. For example, we could say that the serious tone in Metal Gear Solid 3 was good (or at least done well) and the humorous tone in Metal Gear Solid 4 was not (or at least it was not done very well). Pretty much what you said later in your comment, and what I'm gonna allude to later in the second part.
It's ironic, really. My main argument is that we should remove historical context from quality discussion of games, as "seminal" has been forgotten and replaced with "classic." Games like Defender and The Ocarina of Time are regarded as "classics" because of their historical place and influence rather than their longstanding quality, which I'd argue are both limited.
I'd probably argue for that, too, if I had thought to relate it to the rest of this work. Basically, good games don't rely on context to make them good; they make their contexts irrelevant. The problem, though, arises with modern games; since we're in the context for which it was designed (or, rather, a context for which it was designed), it may be difficult to remove ourselves from it. Sadly, the only solution I can proffer is that we should hone our critical capabilities.
However, my answer to the problem of the child is much more simple than that; one should definitively differentiate between the things that they like and the thing they consider "good." For (an extreme) example, receive more glee and learn more from The Room and Birdemic: Shock and Terror than I do from the far superior Escape from New York. More tastefully, I derive far more from the experience of listening to Rihanna than I do from the far more talented and personally representative Rush because I am allergic to their style and singer more than necessary. If Pokemon Black is "worse" than Pokemon Blue, it is because the balancing of the gyms, the roster of Pokemon, and the structure of the journey aren't as strong as that original Pokemon game; however, I'd imagine it's probably improved, and I'm simply a cantankerous old bastard who likes what I like. Analysis of quality involves more objectiveness and decontextualization than subjective analysis, but I always remind people that I think it's okay to like bad things and that I do so regularly.
I made that argument early in the comments (the "separate like and good" thing), so all I can do is nod and post this GIF.
Oh, and agree with the last two people who commented here, @churrific included.