Smyth 2533 inverse attraction

Hello!
While trying to understand how the attraction works I stumbled upon this paragraph of Smyth:

[*] 2533. Inverse Attraction.—An antecedent nominative or (oftener) accusative may be attracted to the case of the relative. The attracted antecedent is often prefixed for emphasis to the relative clause, which thus separates it from the verb it governs or by which it is governed. Cp. urbem quam statuo vestra est, and “Him (= he whom) I accuse, By this, the city ports hath enter’d” (Shakespeare), where the antecedent is attracted into the case of the (omitted) relative.

τάσδε (for αἵδε) δ᾽ ἅσπερ εἰσορᾷς . . . χωροῦσι but the women whom thou seest are coming, S. Tr. 283, πολιτείαν (for πολιτεία) οἵαν εἶναι χρή παρὰ μόνοις ἡμῖν ἐστιν we alone have an ideal constitution (lit. such as ought to be).

I wanted to ask about the last sentence from Smyth’s example. I do not see quite clearly why in this case πολιτείαν isn’t a regular form and Smyth wants to see there nominative πολιτεία which was attracted? For me, as it is dependent on εἶναι χρή, it is natural that it should be in accusative as a subject of an infinitive.

Hi, you may find the Cambridge grammar’s explanation of this same example more helpful (sec. 50.14):

πολιτείαν is subject of ἐστίν, so the expected case for it would be nominative; instead, it attracts to the accusative case of οἵαν (acc. with εἶναι χρή).

πολιτείαν does not sit within the relative clause.

Try mentally replacing the relative clause οἵαν εἶναι χρή with a modifier, and you’ll see πολιτείαν here has been attracted into the case of οἵαν.

Cheers, Chad

Thank you, Chad! After the explanation it became quite obvious.