For example, buildings are made out of the materials we make them out of because we choose materials that won't collapse,

But that simply isn't an adequate explanation for why we build materials out of the materials we make them out of.

This can be easily demonstrated by listing the materials that go into the construction of any building in your town and ranking those materials by how resistant to collapse they are. You will find that lots of the materials involved -- glass, gypsum board, fiberglass insulation, etc, etc, etc. -- are not resistant to collapse; we pick those materials for other reasons.

Of course, someone might complain that I'm being unfair here... of course "because they won't collapse" isn't an explanation of why we choose all the materials we build buildings with, just why we choose sturdy materials like wood or metal or cement to build the framework of a building. Duh.

And, well, yes, this is precisely my point. If I want to usefully answer a question like "Why do we make buildings out of the materials we make them out of?" the most satisfying way to do so is to break it down into more specific questions. "We build frameworks out of these materials because..." "We build windows out of these materials because..." "We build interior walls out of these materials because..." and so forth.

And if someone interrupts us to impatiently say "No, no, no, I don't want to hear about frameworks and windows and interior walls, I asked about buildings!!! I want to know why we build buildings out of the materials we build them out of!!! All parts of a building!!!" all we can really do is encourage them to be less impatient, because we can't usefully answer the question the way they insist on having it answered.

Of course, someone smarter than me might be able to skip the specific-questions stage altogether and construct such a general strategy or itemized explanation solely by analyzing the general question... but if I'm not that smart, I'm not that smart.
Apparently I'm even less smart, because I have no idea what you're saying here :(

Well, for example, suppose Sam asks me "how do I choose what kinds of wine go with what kinds of main course?" I might reply "Well, if I'm serving beef, I serve these wines, and if I'm serving fish, I serve those wines," etc etc etc.

Then Sam, who is much much smarter than me, looks at all of that and goes "Oh! I see. The general rule is to calculate at the ratio between the alcohol content and the tannin content expressed in these units, take that number mod 7, then take the numeric equivalent of the first letter of the main ingredient of the dish in German and take that mod 7, and match wines to dishes based on those two matching numbers."

And I stare incredulously at Sam, perform that calculation for a bunch of main dishes and wines, and ultimately say "Holy crap! You're right!" And Sam says "Of course. Say, why did you choose to answer the question in such a piecemeal way? It seems inefficient."

To which the only answer I can give is "Because I'm not nearly as smart as you are, Sam."

Right, but I'm asking for an explanation of that whole stack of turtles - not how an individual turtle stays up, or even how every turtle stays up, but why the universe is not in the counterfactual no-turtle state.

Right.

And again, my only choices when answering such a general question are:

(1) Be uselessly general ("Things as they are are as they are and have been and will be that way because they got that way because things were as they were in the first place!")
(2) Approach the general question by breaking it down into specifics ("Turtle #1 stays up because of Turtle #2. Turtle #2 stays up because of Turtle #3. Etc."; "We build frameworks out of these materials and windows out of _these materials and etc,")
(3) Be really really smart and come up with a general explanation ("The way you pick a wine is...")

(Of course, another option is the one you adopted for buildings: "we choose materials that won't collapse". Which has the unfortunate defect of being false. The same has historically been true of general theories of human psychology. But, sure, someone really really smart could articulate an accurate general theory here, as per approach #3.)

#3 is obviously preferable, but if I'm not smart enough to do it, I'm not smart enough to do it, in which case #2 is usually my best option.

And if someone impatiently says "No, no, no, I don't want to hear about turtle 1 and turtle 2 and turtle 3, I asked about the stack of turtles!!! I want to know how the stack stays up!!! The whole stack!!!" all I can really do is encourage them to be less impatient, because I can't usefully answer the question the way they insist on having it answered.