W3C home > Mailing lists > Public > public-esw-thes@w3.org > February 2007

Re: Exactly what does broader/narrower mean?

From: Leonard Will <L.Will@willpowerinfo.co.uk>
Date: Mon, 19 Feb 2007 18:15:59 +0000
Message-ID: <+5rt7WOfle2FFArg@mail.willpowerinfo.co.uk>
To: public-esw-thes@w3.org

In message <000c01c75425$f53f2db0$0300000a@DELL> on Mon, 19 Feb 2007,
Stella Dextre Clarke <sdclarke@lukehouse.demon.co.uk> wrote
>If the vocabulary is a taxonomy or a classification scheme, then the rules of
>hierarchy tend to be much looser. It is convenient in a library or a
>bookshop to be able to browse through, for example, the Chemistry
>shelves to find books about chemical elements, chemical laboratories and
>even chemists. (Read more about this in the forthcoming BS 8723-3.)
>
>So to use your example, a thesaurus should not allow German politicians
>to be treated as a narrower term of Germany (although conceivably these
>could be considered as related terms). But a taxonomy or a classification
>scheme could allow you to organise these classes hierarchically if you
>wished.

I have a slightly different approach to this, which probably depends on
a different interpretation of the word "hierarchy".

I agree with Stella that the broader/narrower relationship should be
restricted in the ways she says are specified in BS8723-2.

A "hierarchy" as shown in a classification scheme may or may not be an
example of broader/narrower relationships, depending on whether the two
linked terms belong to the same facet or not. Thus

politicians
    prime ministers

or

Germany
    Bavaria

are valid BT/NT relationships, because in both cases the terms come from
the same facet, "people" or "places" respectively.

The relationship

Germany
    politicians

is not a BT/NT relationship, because the terms come from different
facets. In this case we are expressing a compound concept, "Germany :
politicians".

It is clear that the relationship is not hierarchical, because it would
be perfectly valid to express it as

politicians
    Germany

if you prefer the citation order of facets "people - place" rather than
"place - people".

The point is that a classification scheme provides for combination or
"pre-coordination" of concepts in this way. A thesaurus generally deals
with the elementary concepts, for use in post-coordinate systems where
concepts are combined at the time of searching.

There is a subtle complication in the case cited of "German
politicians", which is not the same concept as "politicians in Germany".
In a classification scheme this could be expressed as arrays within the
"people" facet, with two criteria of division, where the hierarchical
relationships are indeed valid BT/NT:

people
    (people by occupation)
    journalists
    politicians
        (politicians by nationality)
        French politicians
        German politicians
etc.

OR

people
    (people by nationality)
    French people
    German people
        (German people by occupation)
        German politicians
etc.

The node labels that I have shown in parentheses are often omitted, but
including them makes the structure clearer.

In the example of chemical elements, a classification may well specify
these under a discipline such as "chemistry". Again we have compound
concepts, combining terms from a "discipline" facet with terms from a
"materials" facet. Chemical elements may well occur combined with other
disciplines, such as "jewellery - gold" or "fuel technology - hydrogen".
In none of these cases does the combination imply a BT/NT relationship
between the component concepts.

I don't think that SKOS has yet been developed to deal fully with
compound concepts, but I do think it is important that these compounds
should be distinguished from BT/NT relationships.

Leonard Will

-- 
Willpower Information       (Partners: Dr Leonard D Will, Sheena E Will)
Information Management Consultants              Tel: +44 (0)20 8372 0092
27 Calshot Way, Enfield, Middlesex EN2 7BQ, UK. Fax: +44 (0)870 051 7276
L.Will@Willpowerinfo.co.uk               Sheena.Will@Willpowerinfo.co.uk
---------------- <URL:http://www.willpowerinfo.co.uk/> -----------------
Received on Monday, 19 February 2007 22:28:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:55 GMT