SignBank spelling order

Sat Oct 12 11:32:39 UTC 2002

An interesting point, Ingvild, and I heartily agree.

Signs, when they are finished, are a gestalt, and it
is sometimes hard to pick out what features are most
"prominent".

For the Libras Dictionary project, we ran into the
same difficulty.

The system we hit on, and are currently working with
follows the SSS systematically, so that at least we
are working off the same page.

This system is pictorially shown in the Sign Writer
Journal link on the SignWriting.org Web site (first
article).

This system is based on the appearance of the whole
sign, not a linguistic "meaning".

1) Start with the primary hand.  (I know that this
sometimes may be difficult to judge, particularly when
a sign may have both hands moving together, but choose
the "orientation" of your system systematically.  Are
you going to "uniformly" use "right hand" as your
primary hand, or "left hand", or "both".  Choose one
and stick with it, at least for a first draft of your
dictionary."  Using this as STEP 1, order all the
signs by primary hand.  (If you have signs that have
no hands--like shoulder shrugs, tongue movements, or
the like--put them aside for later)

2) What is the orientation of the primary hand?  This
will order all the signs with a given primary hand
into the six orientations and 8 rotations (at least in
the electronic format) fairly quickly (that's 96
ordering points).

3) Is the primary hand articulating (as opposed to
moving about).  If so, follow the order of the SSS for
articulations.

4) Is the primary hand touching the body or located
near a body part that must be shown on the sign?  If
so, start from the top of the head to the feet and
from left to right on the body.  With multiple
articulation points on the head, neck, right arm,
body, left arm, and legs, this also breaks out signs
into a logical order.

5) Is there movement with the primary hand?  For now,
follow the SSS for the order of movements, using this
to be your next step in the system.

6) Does the hand change shape to a secondary position?
 If so, again follow the SSS for the secondary
handshape.  A fingerspelled sign, for example, would
logically follow the first handshape and orientation,
and then on down the line.  If one is ordering purely
by handshape so that the system is based on the hands
themselves, not by comparison to the spoken language
being articulated and the order of handshapes in that
system, one will get differences between an A-B-C for
example, and the order of the handshapes in the SSS
(B-C-A).

7) Is there a second hand? If so, follow the above
sequence 1-6 for the second hand.

8) Is there a second articulation position (like a
sign where the two hands are far apart, and end being
together).  Start with the beginning of the movement,
order by that position, and finish with the end of the
movement, ordering by that position.

As Marianne Stumpf and I ordered the 1600 signs in our
Libras dictionary, each of the signs, if viewed as a
whole, fell into a reasonably sequenced order.

Ordering the signs this way enabled our several
researchers to follow the sequence logically, without
having to worry about how the signs were entered into
the database, or to have to have a linguistic meaning
attached.  As a person unfamiliar with Libras when I
began, I was able to order the signs logically without
having to have a native Brasilian signer at hand.  A
person entering the database may systematically view a
sign any way he or she desires, but at the end, when
the sign is finished, a recognizable spelling is the
same, even if one starts with the secondary hand in
entry or even the point of articulation.

Our system works for us.  We encourage others to
compare, contrast, and give us feedback.

Valerie has created an open-ended system and
encourages experimentation.

--- Ingvild Roald <ingvild.roald at STATPED.NO> wrote:
> Hi all,
>
> as I've started entering signs and their spellings
> into the bank, it
> becomes apparent to me that as we are creating the
> first multilanguage
> sign dictionaries in the world, and they are
> supposed to be working
> together, we had better come up with some rules
> about how we enter the
> spellings. As I understand, that will affect the way
> the SignBank will
> sort and look up by symbols.
>
> It seems (I'm at home now, it's Saturday, and I have
> the refererence at
> work) that signs in themselves are best described as
> one segment each. In
> spoken language, you take the first sound and then
> the second, and so on,
> in a multisegmental way. Even if time plays a part
> in Sign production, it
> is the whole sign that is the sign, and a movement
> is a movement from  or
> a movement to, as the case may be.
>
> Let me try to illustrate by a few signs, attached:
>
> absolutt-01-NO  can be translated into English as
> absolute, bu also as
> absolutely.  I eneterd the first handshape, and then
> the strike, because
> that is what distinguishes this sign from the
> 'finnished' ('has' done ..).
>
> Then I entered the downward, and lastly the passive
> hand.
>
> But then I had to enter the next sign ,
> absolutt-02-NO, which can best be
> translated as absolutely, really, true. Both hands
> are symmetrical, and
> the striking is essential, the fact that it will
> come about by the right
> hand moving down on the left is not of importance
> and not written.
>
> (''noyaktig' schoud be written whit our Norwegian
> o/, and means
> 'accurately, excately')
>
> Then I started thinking: If each and every one of us
> are entering the
> spellings by "importance", it would be hard to get
> an agreed-upon way of
> ordering the spellings. And as the signs should
> probably be seen as whole
> thimgs, the way we pull them apart by arranging the
> writing symbols should
> not be of importance to the understanding of the
> language, but very
> important to the making of dictionaries that work
> across language borders.
>
> So what rules should we follow?
>
> I know Valerie has addessed this question breifly
> before, but now as we
> sit down and actually do this, we should have a
> discussion and make rules
> that will also be felt as appropriate by the sign
> language lingiusts and
> by the signing communities, as far as that is
> possible.
>
> So I dare not enter anything more at the moment ...
>
> Ingvild
>
>

> ATTACHMENT part 2 image/x-ms-bmp
name=absolutt-01-NO.bmp

> ATTACHMENT part 3 image/x-ms-bmp
name=absolutt-02-NO.bmp

> ATTACHMENT part 4 image/x-ms-bmp
name=noyaktig-01-NO.bmp

__________________________________________________
Do you Yahoo!?
Faith Hill - Exclusive Performances, Videos & More
http://faith.yahoo.com