Right then. Plainly, it seems then that the overwhelming order in related scripts is Brahmic -- letters are ganged up in groups, palatals, labials, whatnot. For examples, from the non-Javanese scripts Batak, Rejang, Kerinci, Lampung, and Bugis. I've filtered out many of the letters Baybayin doesn't have, but kept the remaining letters in order:
a(from h) h(from k) n m t d l r b p w y j s g ng ny i u
a(from h) h k b p n w g j d r m t s y ng l c nd mb i u
k g ng t d n p b m c j ny s r l y w h a
k g ng t d n p b m c j ny s r l w y h a(from h)
k g ng p b m t d n c j ny y a l r s w h
k g ng p b m t d n c j ny y r l w s a h
k g ng p b m t d n c j ny r l w s h a
k g ng p b m t d n c j ny r l w s a h y
k g ng p b m t d n c j ny y s w l (a?)
I've split the latter 7 into groups by sound-correspondence. And I'm going to try to analyze them in groups.
The first group is easiest, because all 7 start with the k-series: (K G NG). This group is present in Baybayin, and I hypothesize that Baybayin collation may have begun with K G NG.
The next three groups are nearly as easy, referring to labials (P B M) and dentals (T D N) and some sort of ts- or ch- series (C J NY NC). The C-series isn't in Baybayin. The other two are, but which order to use?
As a more remote cross-reference, here is Kannada, the precursor to Telegu, and the Granthas, the script used since the 5th century and a precursor to a lot of scripts in India. In both cases, I've excluded most of the letters that don't pertain to Baybayin:
K G NG C J Ny T D N P B M Y R L V Sh S H
K G NG C J Ny T D N P B M Y R L V Sh S H
So in other words, the older or more conservative scripts have T D N first, while the younger ones place P B M in front. Either way would be fine. For now I will collate P B M before T D N.
Now what to do with the remaining letters?
Granthas and Kannada somewhat loosely group them into [Y R L V] and [Sh S H].
I see R-L-W and Y-R-L-W. There's also a fair case for following with S, then H, then the vowels. But there's no compelling evidence to say if Baybayin actually followed the order of one of these, or went its own route. My preferred order would be traditional. And would probably be wrong.
So my hypothesis, for what it's worth, is:
Baybayin collation order hypothesis, mostly unsupported:
K G NG P B M T D/R N Y L W S H A I U
In grid form:
K G NG
P (F) B (V) M
T (Th) D/R N
(Ch J Ñ)
Y (R) L W
(Sh) S H
A I U
Note that I've included slots for potential, non-existent F, Th, Ch, J, Ny(Ñ), V, and Sh characters, according to their places in related scripts.
