In his 4th question, Peter Galison asks about the manufactured fundamental; essentially, what do we mean by fundamental? How does something get to be fundamental, and if we manufacture a new fundamental (like, say, a transuranic element) is it natural or artificial?
This seems like sort of a gimme. In one sense, it’s a very old question. Many seventeenth-century objections to experimental knowledge hinged on the fact that an artificial manipulation of nature did not represent a form of knowledge that could be considered natural, and eternal, and, thus, philosophically interesting. Bacon and Galileo argued otherwise, and, over the course of the century, opinion slanted their way. So, really, engineered states of nature have always been a part of modern scientific inquiry. If lasers, or engineered proteins, or quark interactions, or nano-scale technologies seem artificial, I don’t see anyone really seriously objecting that they are worthy of study as objects of interest, and whether one chooses to view them as natural, artificial, fundamental, or whatever, strikes me as a matter of linguistic convention, perhaps worthy of Scholastic debate. They are constrained states of nature, like everything else.
A somewhat more difficult question is how do things come to be fundamental? This plays right into the old SSK questions. How are quarks constructed? Or, following the rival approach, how do experiments end? The old approach seems to emphasize the social resolution of conflicting opinions in controversy. They are replete with the politically defeated scientists, often sulking into retirement stubbornly clinging to their positions. But is that really typical? Don’t scientists willingly change their minds all the time? I don’t deny that the processes of convincing and being convinced are based on social-linguistic traditions, but, at the same time, the role of evidence is clearly important. Here’s my opinion: the future of historical study in this area will focus on robustness. At some point the evidence fits so tightly together that you feel compelled to acknowledge its persuasiveness.
At some point after 1900, even though atoms have been persuasively argued for for a century, you really should start admitting that they exist. The distinction between whether you feel they exist ontologically or phenomenologically ceases to matter, because your practices with respect to them will be the same in either case. To put a little sauce on it, take the quantized field: do virtual particles exist? The name “virtual” even acknowledges that they have a quasi-ontological status (rooted in the superposition of discrete quantum states), but, for all intents and purposes, the robustness of their use in HEP theory effectively secures their reality. Whether they are what we (and, by we, I mean physicists) think they are (if, indeed, physicists think about it at all) is another matter.
Exactly what “robustness” entails has social and physical elements that are definitely worthy of study. Actually, there’s a specialty in operations research dedicated to the idea of robustness–might be worth checking out the technical literature on it.