Where people mark on a map for go to Los Angeles from San Diego, he may
1. cirle Los Angeles, then circle San Diego, then draw an arrow; or
2. draw an arrow, then circle Los Angeles, then circle San Diego
in other words, the order can be O1 O2 V or V O1 O2, the order of objects are fixed, but they may be before or after Verb.
in this case, how can I define the grammar? I think about the "|", but it only control one arguement. Am I right?
Thanks,
Yong
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Set categories themselves do not change the power of the system, but one can define rules which increase power, such as the ones that Beryl Hoffman proposed. See chapter 6 (and in particular p 155-156) of my dissertation for discussion:
The way I use sets does *not* change the power -- they basically serve as schemata over several category instantiations. FWIW, using multiple categories is a fine strategy in many situations.
As for your question on assigning semantics, it is too vaguely stated for me to help you out. Perhaps it would help you to look at the rough guide that Cem, Mike and Geert-Jan put together:
Where people mark on a map for go to Los Angeles from San Diego, he may
1. cirle Los Angeles, then circle San Diego, then draw an arrow; or
2. draw an arrow, then circle Los Angeles, then circle San Diego
in other words, the order can be O1 O2 V or V O1 O2, the order of objects are fixed, but they may be before or after Verb.
in this case, how can I define the grammar? I think about the "|", but it only control one arguement. Am I right?
Thanks,
Yong
You need two categories, one for the verb-initial frame, one for the verb-final one.
Can it be done in one category? like "|". thanks
Yes, with set categories: s|{np1, np2}. However, note that this will also give you other orders like np1 np2 v, np2 np1 v, v np1 np2, and v np2 np1.
Jason
Thank you, Jason,
in other words, s|{np1, np2} can produce 4 orders. In that case, how can I assign the semantics according to "from" and "to" within openccg?
Thanks
by the way, does s|{np1, np2} have any advance over context-free definition? Just in grammar representation, or computational complexity as well?
Set categories themselves do not change the power of the system, but one can define rules which increase power, such as the ones that Beryl Hoffman proposed. See chapter 6 (and in particular p 155-156) of my dissertation for discussion:
http://comp.ling.utexas.edu/jbaldrid/papers/dissertation.html
The way I use sets does *not* change the power -- they basically serve as schemata over several category instantiations. FWIW, using multiple categories is a fine strategy in many situations.
As for your question on assigning semantics, it is too vaguely stated for me to help you out. Perhaps it would help you to look at the rough guide that Cem, Mike and Geert-Jan put together:
http://comp.ling.utexas.edu/wiki/lib/exe/fetch.php/openccg/openccg-guide.pdf
And also look at the tutorial materials for OpenCCG that we have developed at UT Austin:
http://comp.ling.utexas.edu/wiki/doku.php/openccg
Jason