Thanks Chad. Smyth actually discusses this passage in a section on attraction and writes:
μὴ ὑποκειμένων οἵων δεῖ θεμελίων (for τοιούτων οἷα δεῖ ὑποκεῖσθαι) if the foundations were not as they ought to behttp://www.perseus.tufts.edu/hopper/tex ... thp%3D2522
I guess that squares with your reading, but my head hurts now. I guess my question is why Smyth puts a ὑποκεῖσθαι in the relative clause. I understand why you do it I think. But Smyth's seems gratuitous as unpacking the attraction doesn't seem to demand it. Shouldn't Smyth have just written: μὴ ὑποκειμένων οἵων
(for τοιούτων οἷα) δεῖ θεμελίων? On this reading, δεῖ is all there is inside the relative clause. If look at the Xenophon, isn't that all there is to the relative clause? δεῖ? And if we unpack the attraction, isn't that all there is? δεῖ? And to approach the question from another direction, in Xenophon, post attraction, where is ὑποκειμένων? Outside or inside the relative clause? And pre-attraction(=sans attraction=unpacked), where was it? Outside or inside the relative clause? And if it has moved, how is that possible?! "Using" attraction never requires changing more than the demonstrative and the rel. pronoun does it?!