Observation about impropriety

Consider an MOS scale. Rothenberg predicts an increase in some sort of
"confusion" which occurs with the intelligibility of melodies,
particularly when transposed, if you tune this MOS to an "improper"
variant (L/s > 2/1) vs a proper one. Let's call this "impropriety
confusion." (Let's just not consider scales with only one "s" interval
per period right now, which can never be improper no matter how you
tune them.)

Given that, I'm curious how many of you agree with this statement: For
any MOS of the form xLys, impropriety confusion is increased as y
increases and x decreases; e.g. improper 2L5s is more noticeable than
improper 3L4s is more noticeable than improper 4L3s is more noticeable
than improper 5L2s.

From a mathematical standpoint this seems trivial to me, because the
extremes for any MOS of the form xLys will be x-EDO and (x+y)-EDO. So
if you're at 1L6s, then your options are 1-EDO and 6-EDO - and it's a
long way till you get to 1-EDO. And if you're at 2L5s, your options
are 2-EDO and 7-EDO, and there's a lot of opportunity for impropriety
as you get towards 2-EDO. But if you're at 5L2s, then your options are
5-EDO and 7-EDO, which isn't nearly as bad.

For example, consider magic[7] vs mohajira[7], which are some good
temperaments for improper vs proper 3L4s respectively. To my ears, the
shift between these two is noticeable and dramatic. They barely sound
like the same scale at all. The "major seconds" in magic[7] sound like
there's a "hole" in them or something, they sound like they should be
thirds or fourths or something, comparatively speaking. I can
immediately tell that two "minor seconds" put together forms an
interval much smaller than the "major second." For mohajira[7], this
isn't the case.

-Mike

OK, one more thing that's also important. In the last message, I noted
that the "major seconds" in magic[7], which are 6/5, sound like they
have a "hole" in them, or that there's a space where a whole bunch of
minor seconds could fit.

I also note that in something like 19-equal, the amount of minor
seconds that could fit is definitely more than 2, which means that
there's some sort of super-impropriety in which a perceived chain of
three minor seconds stacked together is even still much smaller than
the major second. Another way to put this is that in 19-equal, not
only is magic[7] improper, but magic[10] is improper as well. My brain
seems to easily pick up on this.

However, I notice the same exact thing for augmented[6] in 12-edo:
those 6/5 "major seconds" seem to take a huge amount of space up, such
that my brain can easily tell that more than two minor seconds fits
into the space of each major second. This scale, however, is proper.
On the other hand, both 5L2s and 6L1s sound more or less proper to me
unless you tune them so severely that s almost vanishes. That is, I
perceive the proper 3L3s as being more analogous to the improper 3L4s
than either proper 5L2s or improper 5L2s.

I'm curious if this perception is shared by anyone else. -Mike