Bart,
If I understand correctly:
1. The larger physical size of a Medium Format sensor compared to a 35mm format sensor means that a longer focal length lens is required to return the same field of view.
2. By definition the magnification of a longer focal length is greater than a shorter focal length.
3. Greater magnification increases MTF.
Correct. Assuming similar Fields of View, the larger magnified detail is 'lower spatial frequency' detail on the MTF curve and thus has a better preservation of the original subject contrast, and as a consequence more tonality to work with.
4. File bit-depth sets a limit on the tonality that can be captured - a 16-bit file may retain 100% of the tonality inherent in an image (depending on the image content) whereas an 8-bit file may curtail it.
It's the ability to discriminate between subtle tones. A 16-bit/channel file can encode more intricate detail/tonality for post-processing. It's then up to the postprocessing to enhance or reduce those differences, depending on the intended effect.
Is this correct?
Yes, that's my take on it. Of course, as Doug pointed out, there are more factors like lens quality, but when all other factors are more or less the the same, MTF still makes the difference, makes the files more malleable. That's what it boils down to, IMHO.
Cheers,
Bart