Thus... with your reasoning... if you combine two 18" mirrors with a precision greater than 1/4 lambda, they will deliver the resolution of a 25", right?
That is not so easy. If two identical round apertures would be combined by coherent way as interferometer, the Point Spread Function (PSF) of such sintetic aperture looks like Airy disc of a single aperture dissected by a number of parallel lines. As more separation between apertures (so called base of interferometer), desto more and thinner getting lines (or fringes) splitting Airy disc. For polychromatic light the best contrast would be for zero fringe. As fringe scale depends on wavelength (~wavelength/base), contrast of fringes decrease away from zero fringe. Point Spread Function looks like on a picture below. That is how single stars should look like in such sintetic aperture. Any extended object would look as a convolution of PSF over every point of object resulting quite complicated image. So getting high resolution in interferometry is not so straight forward.