Type identity

[Detmar Meurers]

Hi,

> Supposing that
> there is a type-identity checking mechanism in place,
>
>    what's the current situation there, BTW. Does RSRL provide means to
>    express type (in)equality??
>    Berthold


As far as I see, adding type (in)equality to an attribute
value-logic like SRL or RSRL is simple syntactic sugar (and does not
involve a meta-level as Frederik hinted at). The property making it
straightforward to add such type (in)equality syntax is that the
linguistic ontology is explicitly declared in the signature (type
hierarchy + appropriateness conditions), so that we know the finite
number of different types which can occur. The semantics of such a
type identity syntax therefore only has to check a finite set of
explicit possibilities.

Type identity is very useful for formalizing lexical rules in SRL,
where one has to preserve a type value of a path occurring in the
input of the rule if that path isn't specified in the output. More
details can be found in "On expressing lexical generalizations in
HPSG", which just appeared in Nordic Journal of Linguistics 24 (2).

Lieben Gruß,
Detmar

--
Detmar Meurers                              Fax: Int + 614 292-8833
The Ohio State University                   Tel: Int + 614 292-0461
Department of Linguistics                   E-Mail: dm at ling.osu.edu
1712 Neil Avenue, Oxley Hall     Homepage: http://ling.osu.edu/~dm/
Columbus OH 43210-1298, USA    PGP key on web page (use encouraged)