"finalized" signs

[Rachel Channon]

I wrote a paper (1992) about initialized signs in ASL, which  looked at the
judgments of one native signer to about 450 different initialized signs. The
paper concluded as follows: (Underlining indicates the letter of the gloss
used as the handshape).

.....This study shows that the ideal initialized sign would appear to have
the following characteristics:  It is a noun, and preferably a proper noun.
The meaning of the sign is not iconic: it is not easily pointed to or easily
shown by drawing or example. The first letter of the English gloss is used
for initialization and the gloss is 9 or more letters long.  The signer does
not wish to fingerspell it and does not know another sign which he/she
considers equivalent in meaning.  Preferably it is a one-handed sign, made
in space.  If it is a two-handed sign, both hands are initialized,  only one
handshape occurs on either hand, and it is made with the hands in contact
with each other.  The initial handshape is N, I, E, or D.  Some good
examples might be: INSURANCE, INFECTION, EASTER, and DEMOCRACY.
Furthermore, the good initialized sign watches its company, and does not
appear with too many of its fellows.
The worst initialized signs have the following characteristics:
They are verbs, adjectives or pronouns.  The meaning is easily demonstrated
through pointing or drawing.  The first letter of the gloss is not used for
initialization, and the gloss is very short, less than 5 letters.  The
signer prefers to fingerspell it and/or knows another sign considered
equivalent in meaning.  It is a two handed sign where the weak hand is not
initialized, and more than one handshape occurs on either or both hands.
There are multiple locations, such as a sign that moves from the body to
face or body to arm, or it is located on the face.  The initial handshape is
O, W, R or M.  Some examples of unacceptable initialized signs would be
*MARVEL, *ROT, *RICE, *WELL, *MAIL, *ORBIT and *THEY.
Perhaps the most interesting findings were that initialized signs are more
appropriate as nouns, and for naming concepts that are not iconic.  These
signs are at the opposite pole from classifier predicates which are strongly
verbal, spatial, and iconic.  As Padden (1991) remarks, fingerspelling is
also more acceptable for nouns.  This suggests that the influence of the
surrounding hearing, English culture is most easily assimilated or borrowed
from in the specific grammatical category of nouns....

Clearly there are exceptions to each of these tendencies.  For example,
THURSDAY is an initialized ASL sign which has two letters as handshapes, and
one of these is not the first letter of the gloss .  The finalized and
medialized signs that people have been mentioning like TEXAS  are other
examples.  It seems to me that what makes these forms acceptable is that
there is another equally important sign in the same semantic domain (such as
TUESDAY)  which is also initialized and has already grabbed the first
letter.  Both days of the week and names of cities have a pattern in ASL of
being initialized, so there is a bias within these word groups  in favor of
initialization and against homonymy which overcomes the constraints against
medialization and finalization, etc.  (No doubt, one could put this in
formal Optimality Theory terms as well.)
Rachel Channon