He offers a great example of what is known in the philosophy of science as the pragmatics of explanation. You see, all such interesting questions are why-questions, and why-questions are a request for explanation of a specific sort. To adequately respond to why-questions requires that we do a number of things, the least of which is to identify the relevant class of answers and contrasting responses. I believe it was Bas van Fraassen in his famous work The Scientific Image uses the example of Eve in the Garden eating the apple. Why, we might ask, does Eve eat the apple? Well, this is an interesting question on a number of fronts. We could be saying why was it Eve that ate the apple, in contrast to someone else eating it. We could ask why was it that Eve ate the apple, in contrast to doing something else with the apple. We might also ask why was it the apple the Eve ate, in contrast to some other fruit. So asking one question can be regarded in these three examples as three entirely different questions depending on the contrasting class. By contrasting, we can identify what is a reasonable answer. Feynman goes the extra step in his examples of identifying the fact all answers come within a framework because to meaningfully talk about the concepts involve require that the person being answered can understand the terms involved. This has consequences to an adequate answer to a question. If someone asks how many people there are in China, an answer to school children might be “a lot” or “a billion.” The first answer would be sort of unqualified, being that a lot compared to India would not be accurate, but it does get the picture across nonetheless. The second answer is numerically false as there is over a billion, but it still serves the purpose of explanation. Now if we were in a higher education class that requires more exacting demographics, those answers are not satisfactory. We would require something along the lines of “1.3 billion in 2008 as cited by the World Bank development indicators.” So a meaningful answer to a why-question requires recognition of the contrast class of answers as well as a recognition of what sort of satisfactory response the audience can be expected to understand. The example of demographics I give, however, is quite straightforward. The issue Feynman talks about, with both the broken hip and magnetic repulsion, has even deeper consequences involving linguistics. As he points out, if an alien came to the planet they would have to understand something about the organization of human affairs, such as husbands calling for ambulances when their wives are hurt or taking people to hospitals when they are hurt. But when we get to real hardcore scientific abstractions like forces and fields that we cannot directly observe, we have to admit ignorance. This is the beautiful part of Feynman’s closing words. What we have to accept is what we observe. We observe these pieces of metal, oriented in certain ways, repelling and, when reversed, attracting each other. This is the fact of the matter. Our explanation of it depends on our framework and that is provided by our model of the phenomena. The model is only a representation. We invent this idea of an electric force that behaves in certain ways that satisfy our observations. Now, why we pick this theory against others depends on many things that go far beyond the issue of explanation, but that is precisely what is being talked about here. But this can be misleading, because notice how he talks about the force of gravity. Is gravity really a force like the others? It doesn’t seem that way on, say, Einstein’s model of the universe where gravity is the effect on accelerating bodies moving in straight lines in a warped geometry, a spacetime misshapen due to the mass of objects. So gravity is a result of curved space, not a force acting on bodies in what we understand a force to be. But Newton’s explanation of gravity qua force may be just as satisfactory in some instances (e.g., falling bodies on Earth). Interesting stuff indeed.