On 9/11/2012 10:27 AM, doug foxvog wrote:
> The slots (SUBJ, OBJ, BENF, date-1) are very generic, evidently relying
> upon the type of the instance (a MOVE) to give them more specific
> semantics. William's suggested slots (giver, given, receiver) are less
> generic, and better suited to the verb "give". I would prefer using more
> specific relations on encoding the meaning of the statement (performedBy,
> providerOfMotiveForce, fromPossesser, toPossesser, primaryObjectMoving,
> objectOfPossessionTransfer, dateOfEvent), recognizing that if one is
> parsing NL, the more generic ones are necessary (01)
The issue of slot names (also representable as binary relations) has
been treated in great detail in linguistics. There are three common
terms in the literature: case relations, thematic roles, and actants. (02)
The word 'case' is the oldest word, which was introduced as 'casus'
by the Latin grammarians. It was popularized by Fillmore's case
grammar. The word 'thematic' was introduced by Chomsky to avoid
requiring him to cite Fillmore. The Chomsky tradition also uses the
term theta-role (with 'theta' spelled by the Greek letter). (03)
The word 'actant' was introduced by Lucien Tesnière for his pioneering
work on dependency grammar. It is widely used in Europe, especially
for those who follow Tesnière's tradition. (04)
Whatever tradition is adopted, linguists distinguish *syntactic*
relations, such as Subject and Object, from *semantic* relations,
such as Agent, Patient, Theme, Instrument, Beneficiary, Experiencer,
Recipient, etc. When you switch from active to passive voice, the
syntactic roles change, but the semantic roles remain unchanged.
The dative case in many languages can be distinguished as a
Recipient or a Beneficiary. And the possessive or genitive case
has many different roles. (05)
Furthermore, there is a big difference between an active agent,
who does something with a verb like 'eat', from a stative agent,
who experiences something with a verb like 'see' or 'own'. (06)
There is also a big difference between a patient that is changed
by an active verb like 'eat' vs. a theme that does not undergo
any change by a verb like 'see', 'move', or 'own'. (07)
The tradition of using verb-specific terminology, such as giver, given,
and receiver is also widely supported in computational linguistics.
A convenient way to relate these terms is to develop an ontology of
dyadic relations. Then you can specify Driver and Giver as subtypes
of Agent. You can be as detailed as you like and specify TaxiDriver
and BusDriver as subtypes of Driver. (08)
For a diagram that shows such a hierarchy, see (09)
http://www.jfsowa.com/ontology/themes.gif (010)
For a sample ontology with a hierarchy of relations, see (011)
http://www.jfsowa.com/ontology/roles.htm
Roles and Relations (012)
For more detail about thematic roles, see (013)
http://www.jfsowa.com/ontology/thematic.htm (014)
John (015)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (016)
|
