some more comments

0 views
Skip to first unread message

Damir Cavar

unread,
Nov 4, 2008, 5:01:39 AM11/4/08
to GOLD Ontology, lon...@ihjj.hr
Comparing MULTEXT and GOLD, this are some of the issues and comments
(and questions), take this just as a brief brainstorming result:

- Verb Type in MULTEXT: main, auxiliary, modal, copula; there is no
straightforward counterpart in GOLD, maybe it is not necessary
- Adjective Type: qualificative, possessive (and ordinal); not in
GOLD, maybe not necessary
- what about comparative and superlative of adverbs?! (in GOLD only
for adjectives)
- InterrogativeProform as interrogative pronoun; should be OK
- Adverb Type in MULTEXT: general; this is the only type relevant in
Croatian (maybe?)
- translation for Adposition in Croatian missing, we'll adopt
"Adpozicija" for the time being
- Number Type multiple and special from MULTEXT not in GOLD (?)
- Residual and Abbrevation not in GOLD
- Particle Types not in GOLD (maybe interrogative operator only)
- no Neuter in GOLD
- Comitative case as instrumental? (Instrumental morphologically
covers both in Croatian)
- Locational as Locative; needs to be checked!
- can Animate stay in its taxonomic position like that? (should this
be a semantic property, rather than a morphosyntactic case property?)
- Aspect doesn't exist in MULTEXT for Croatian!
- Progressive is Imperfective Aspect in MULTEXT
- Conditional as Modality in GOLD?!
- no Infinitive in GOLD
- Participle as SyntacticWord in GOLD???
- What is the difference between perfect, aorist and imperfect in
GOLD???
- 2nd Futur not in MULTEXT
- do we need Positive as a Polarity? (in Croatian we might!)
- Definiteness and Specificity not in GOLD (is necessary elsewhere,
but for Croatian in particular)
- Numeral Form not in GOLD, but in MULTEXT

Should "ambiguity" be a concept or property? Maybe in the semantic
part (lexical ambiguity, structural ambiguity)??


best wishes
Maja & Damir

Helen

unread,
Nov 15, 2008, 10:58:37 AM11/15/08
to GOLD Ontology
Thanks for this list, Damir and Maja,

I still don't know the answer to your initial question: what do we do
about these suggestions? But I've nudged the people that may have the
answers. We are trying to map a Fufulde lexicon to GOLD right now in
the LEGO project, so I imagine we will have some of the same kinds of
questions very soon.

What I remember from the GOLD discussions is that
(1) new concepts should be attached in the tree to the lowest node
that they match, so that
"Adjective Type: qualificative, possessive (and ordinal)" would simply
be attached to 'Adjective'
(2) any further specification is, in effect, a COPE

I'm not sure what this actually means in terms of the formalism,
however. My guess is that the RDF/xml document of the Croatian COPE
would include the URI for "Adjective" with additional predicates
specifying that this type of adjective is qualificative and
possessive. And perhaps, because this is a COPE, it would not affect
'core GOLD'--though we would certainly want to publicize it on the
GOLD Community website in order to promote its re-use as a de facto
standard.

I am tossing out this 'know-nothing' formulation in the hope that
someone more knowledgeable will now be moved to correct it. :-)

Best,
-Helen
Reply all
Reply to author
Forward
0 new messages