search

amrisi / amr-guidelines

250 stars 87 forks source link

Role for ordinal numbers? #8

Open nschneid opened 11 years ago

nschneid commented 11 years ago

Ordinal numbers provide a distinctive type of modification and form a special morphosyntactic class in many languages. Do they deserve a special non-core role, e.g. :ord (instead of :mod)?

(p / planet
    :ord 2)

the second planet

In contrast:

(h / hand
    :mod (s / second))

the second hand (of a clock)

We would have to decide what to do about expressions that reference a sequence/ordering, e.g. second in a long line of kings or second to last of the finalists.

uhermjakob commented 11 years ago

I agree that it would make a lot of sense. Something like :ord has been on our mental to-do list for a while.

It would require numerous changes. Estimate: about 200 occurrences of first, second, third, ... The AMR Checker could help identify likely AMRs to be modified.

A few details to be worked out include how to deal with ordinal-like cardinals:

nschneid commented 11 years ago

The line between name and number is not always clear for examples like "platform 7". When there is a specifically established designation for something I'm inclined to view it as more of a name, even if some sort of scale/ordering is used to organize the names.

We are using :li only for explicit list markers in the text, correct? Should we distinguish between textual lists, conceptual lists, and other scales/relations (e.g., time) that might involve ordinals?

Perhaps we ought to come up with something that mirrors AMR's treatment of comparatives and subsets, both of which involve a relation between something and its alternatives or counterparts. (Consider: the 3rd most intelligent of the 9 cats)

kevincrawfordknight commented 11 years ago

I have not completely read everything here, but this seems like a good idea for AMR 1.5.

mgeorgescu commented 11 years ago

What if we have bigger ordinals like 259th - do we spell out the number, even if it's rather long? What about 2345th? In this case, it's quite cumbersome to spell them out.

uhermjakob commented 11 years ago

OK, based on a proposal by Natalie, the AMR workshop participants today adopted this new construction for ordinals:

the second planet

(p / planet
  :ord (o / ordinal-entity :value 2))

our first visit in 10 years

(v / visit-01
  :ARG0 (w / we)
  :ord (o / ordinal-entity 
         :value 1
         :range (t / temporal-quantity 
                  :quant 10
                  :unit (y / year))))

We sat in the preantepenultimate pew.

(s / sit-01
  :ARG1 (w / we)
  :ARG2 (p / pew
          :ord (o / ordinal-entity :value "-4")))
mgeorgescu commented 11 years ago

Shall we then start using this and attempt to correct instances of :ord for the July delivery? Or shall we attempt to correct these instances after the July delivery for the September one?

We would be more in favor of the latter - but please advise. We probably cannot do it by July 24. Or are you planning to attempt an automatic correction?

kevincrawfordknight commented 11 years ago

madalina, we're not implementing the new :ord for this release. kevin

On 7/22/2013 12:22 AM, mgeorgescu wrote:

Shall we then start using this and attempt to correct instances of :ord for the July delivery? Or shall we attempt to correct these instances after the July delivery for the September one?

We would be more in favor of the latter - but please advise. We probably cannot do it by July 24. Or are you planning to attempt an automatic correction?

— Reply to this email directly or view it on GitHub https://github.com/kevincrawfordknight/amr-guidelines/issues/8#issuecomment-21328206.Web Bug from https://github.com/notifications/beacon/rwKnZecEA4TCfSMrcy1ztZZPopFv5tl092Ij9JBRq0FOb9iCAZBEIgRnDuPBQOvc.gif

cbonial commented 11 years ago

Now that we have this issue settled with the ordinal proposal, could we have a place in the guidelines where the full options and definitions of each item within ord are defined? For example:

:ord ordinal-entity :value value-interval (definition of what this is) :op1 (bottom end of interval) :op2 (top end of interval) :range (definition of what this is) :quant

uhermjakob commented 11 years ago

I just greatly expanded the AMR Editor help page for ordinals: http://www.isi.edu/~ulf/amr/lib/popup/ordinal.html