Very profound. Are we talking philosophy here, or typography?

I think anything has both practical and theoretical dimensions, and the latter instinctively veers towards philosophy!

hhp

As usual when you cite the 'subconscious', my response is 'Prove it'. But I'll let you off this time, because I think it is more important to cite Pelli et al 2002 [PDF], which to my mind is the most important study yet done in its contribution to our understanding of 'even colour' (I prefer the term even texture) and its functional rôle in reading. Critically, the concept of spatial frequency channels defines limits within which variations in typographic texture may operate without impeding reading by forcing channel retuning. This statement from the study's conclusion is worth re-reading several times:

For each alphabet, font, and size, letter identification is mediated by a single one-or-two-octave-wide visual channel. Even in the presence of very asymmetric noise distributions, observers continue using the same channel to identify and detect letters and gratings. Together, these two findings say that the channel is selected bottom-up by the signal, not top-down by the observer.

There are some simple perceptual tests than one can do to confirm that tuning from one spatial frequency channel to another takes time. Ergo, if text involves too great a span of spatial frequencies, reading will be impeded by the need to re-tune. So, it seems to me that even texture in the sense of letters occupying the same spatial frequency channel, is functionally beneficial to reading, while your 'breaking of even colour' is so far without demonstrated benefit.