You can't measure a 12MHz signal with a 1MHz PIC. Can't be done. Imagine you can catch tennis balls at the rate of one per second. Now imagine some twit loads up a ball shooter and switches it to 100 shots per second.
If you squirt each signal into a divider for which you know the dividing ratio, e.g. 256, then you will get at the output a 12MHz/256=46875Hz signal, which you can count with a PIC, and multiply the resulting frequency by 256 to calculate the original frequency.
That won't help much with the phase difference. Just playing with ideas, if you XOR the output of the two dividers together that will give a short pulse at each end; if you feed that into an RC network then you might be able to measure the rise and fall times, from which you might be able to deduce the pulse width and therefore the phase difference.