[Corpora] [Corpora-List] Named entities and abstract codes

[John F Sowa]

On 11/3/2014 8:38 AM, Jannik Strötgen wrote:
> "the word 'Named' aims to restrict [Named Entities] to only those
> entities for which one or many rigid designators [...] stands for
> the referent" (Nadeau and Sekine, 2007)

That definition is is inconsistent with the following:

> Depending on why you ask, you probably just want to follow
> Jurafsky & Martin: a named entity is everything "that can be
> referred to with a proper name" (Jurafsky and Martin, 2008: p.761)

I agree that this is the most commonly used definition for NLP.
But the W3C notion of rigid identifier excludes most proper names,
since very few proper names designate a unique individual.

> with 'rigid designators' being something that "in any possible
> world [...] designates the same object" (Kripke, 1980: p.48)

This is a philosophical definition that is inconsistent with
both of the above.

It's interesting to relate those definitions to Leon D's question:

> Are "abstract codes" named entities? For example,
> "The *15:07 train to Sheffield*", "*Flight MH17*",
> "The new *Canon 50D*", "pass me *document 123*"?

'Sheffield' is a proper name of cites in England, Massachusetts,
and Alabama.  It is also the name of a Hotel in Manhattan.

'MH17' and 'Canon 50D' are names of types of entities, not
single instances.

'document 123' could be the name of a type or of a specific
instance.  But it would not usually be called a "proper name"
according to typical English usage.

John

_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora