[sw-l] Sign Variants via SWML

[Sandy Fleming]

Antônio Carlos,

I see what you're asking, I've introduced a bias accidentally from the fact
that I was really only thinking in terms of the signs I had thought of as
examples, and I'd somehow concentrated on handshapes.

As for Dan's email about GIVE - yes, a look at some directional verbs could
be very interesting!

Antônio Carlos wrote:

>   I was understanding your proposal of including variants in SWML as a
> means for taking care of prosodic features (intonation) of sign
> language. Your phrase concerns iconicity.
>
>   The question I have is this: can you (or anybody else) explain the
> difference between the two concepts and tell the way they are related in
> sign languages?
>
>    Could symbol variants take care of both aspects?

Perhaps my attempt at showing how natural the idea was by mentioning
iconicity has backfired!

But, seeing that trying to relate it to linguistic features always raises
difficult problems, I wonder if what I'm really after here is something more
basic and based not on the nature of sign languages, but based simply on
what the SWML itself has to offer?

So we have a symbol expressed something like this:

       <symb x-flop="0" y-flop="0" color="0,0,0">
         <x>78</x>
         <y>57</y>
         <category>01</category>
         <group>05</group>
         <symbnum>007</symbnum>
         <variation>01</variation>
         <fill variant="1">05</fill>
         <fill variant="2">06</fill>
         <rotation>03</rotation>
       </symb>

and, without getting into linguistics at all, the application simply allows
us to express variants to our hearts content for any of the elements in the
<symb> element. You can now create thousands of impossible vatiants (without
making the SWML very much bigger), but in practice, which variants are
actually used in a dictionary will be entirely up to the dictionary editors.

This also solves (or at least ducks!) the problem of Dan's GIVE. There may
be a potential combinatorial explosion, but the dictionaey editors only
include the variants that they feel gives the right balance between
versatility and usability that they want from their dictionary.

I think what I'm aiming for is to provide a simple technical solution to the
fact that when building a dictionary we run across the problem that many
signs do have several variants that we'd like to include. It's less work for
dictionary editors, and provides a simple way for them to keep variant signs
grouped together so that they can make their dictionaries richer without
having to worry about whether putting in lots of variants will give them
housekeeping headaches in future.

The educational advantages to the user and other things I mentioned are
perhaps better just seen as bonuses and we needn't worry about the semantics
of it.

Now imagine a BSL user watching a documentary about marine life. He sees a
clip of a flatfish swimming upside down to sift food from the sand on the
sea bottom. He goes and describes this to his friend, then he goes to enter
some more signs in the dictionary he's building. He loads in "fish" because
he realised while watching the programme that he didn't think to enter a
sign for "flatfish". He enters it as a variant of "fish". Then he realises
that he earlier quite naturally used a sign for
"flatfish-swimming-upside-down". Should he put this in as another variant?
We make it technically possible, whether he thinks it's within the scope of
his dictionary to include such unusual variants is up to him!

If it were me, I'd probably decide that "flatfish" is something that some
signers, especially learners, might not think of, and it's worth entering
into the dictionary to help him see, by loading "fish" and requesting to see
all its variants, that this is an alternative he shouldn't forget about. If
he really wants to show the flatfish upside down he'll know this and can
adjust the orientation, I'd say, so I wouldn't bother with that particular
variant unless I had a remit to make the dictionary as big as I possibly
could!

Sandy