Re: Thoughts on the appropriate relationships between coordinated concepts

On Wed, Dec 2, 2009 at 10:51 AM, Ross Singer <rossfsinger@gmail.com> wrote:

> Hi all, I'm looking for some advice for the best way to model the
> relationships between a skos:Concept that represents a "topic", e.g.:
> http://id.loc.gov/authorities/sh85010030#concept
>
> to a skos:Concept that represents a subdivision to be used in coordination,
> e.g.:
> http://id.loc.gov/authorities/sh00006049#concept
>
> These are describing basically the same thing, but they have different
> uses, different contexts and are not interchangeable.  That being said, it
> would be useful to acknowledge the existence of one from the other and that
> it's semantically talking about the same concept.
>
> Any ideas?  I'm leaning towards skos:closeMatch, but I'm not sure if
> skos:relatedMatch (or, perhaps, some different predicate entirely) is more
> appropriate.
>



The two items are identical and should be merged. It's only very recently
that subdivisions have had entries in the subject file, and these entries
don't carry enough information to generally specify precisely when a
subdivision can be used.

Thinking about this  linguistically sense may help.

In the following sentences, *authorship* is precisely the same lexeme, with
precisely the same meaning.

(1) This book is about authorship.
(2) This book is about authorship of poetry.

Being able to be used as a topical subdivision  is a [+FEATURE] of the
lexeme *authorship.* This feature is available to be used as a  selection
restriction to the various constructions that can be formed as subject
headings.

So the best approach may be to specify a boolean property specifying that
the individual concept can or cannot be used as a free floating topical
subdivision.

Simon