Gallows G characters

["GC" <glenclaston@xxxxxxxxxxxxx>]

One of the problems I?m having in attempting to
order alphabets by probable word is that different
characters keep showing up in the same columns, as
if they?re interchangeable.  I've been examining
each character by form and testing its behavior in
an effort to gain some set of rules for their use.
While there should be no attempt to arbitrarily
reduce the alphabet down to character forms, there
are many points that suggest there aren't nearly
as many characters as first appear to the eye.
The G gallows in its two form is a fine example of
this problem.

The standard gallows exists in two forms, a double
loop and a single loop.  The G gallows exists in
two forms, the double loop and the single loop.
The gallows/picnic therefore exists in four forms,
one for each of the above.  The standard gallows
forms exhibit roughly the same behavior, as does
the four forms of the gallows/picnic.  The G
gallows form is an odd duck, and appears to me to
be a homophone of the standard gallows set.  I?d
really hate to have to incorporate this behavior
into a stroke dependent structure, but in cipher I
can explain it as a homophonic replacement scheme
in the system.  There may be other explanations of
course which might be brought to light by better
minds.

In my sample of 35,003 characters, folios
001r-052v, the G form occurs 424 times,
representing about 1.2% of the total characters.
Out of these 424 occurrences, it appears 105 times
as the line initial character, meaning 24.76% of
the time it is used to begin a line.  The most
interesting thing is that 345 times out of these
424 characters, it occurs ONLY in the first line
of a paragraph.  This is 81.36% of the time this
character form occurs in the first line of a
paragraph.  Now that takes some explaining!

There are two options that come to mind for why it
might appear in the body of the paragraph only 1
in five times.  They may indicate a change in the
system, or there may have simply been room to
write a G form instead of a gallows.  (I prefer
the former.)

Whatever theory that finally works has to take
this into account and explain it.  I?m betting
these are homophones of regular gallows
characters, and their use follows some rule as yet
to be explained.

GC