On our Goodman episode, I start out by trying to give a short explanation of Goodman's "New Riddle of Induction." When we're presented with evidence for a general claim, how do we tell which general claim the evidence is in support of? Goodman contrasts the predicate "green," which we might think we can project to future cases when we see that all current emeralds are green, with "grue," which is defined as green previous to this moment and blue after this moment. He argues that our past observations don't tell us which predicate should be projected into the future; we have to give an explanation why we intuitively want to project green and not grue, even though we haven't yet had an experience running counter to the hypothesis that all emeralds are grue. Giving an account of this is more difficult than you might think, and this essay shows Goodman in full-bore analytic mode: very methodical, but still readable and actually fun if you're into that sort of thing, as opposed to his mathematical philosophy, which I think is no fun under any circumstances.
You can read the essay here or get ahold of Fact, Fiction, and Forecast,
the book that contains it, which is really one continuous essay with a few phases: on counterfactuals, possibility, and then after discussing induction, he gives a positive account for how to solve the problems he's put forth, which involves, as you might guess, reference to convention, but of course it's much more complicated than that.