An ontology for cultural history using Campbell's corpus

Do you have a conversation topic that doesn't seem to fit any of the other conversations? Here is where we discuss ANYTHING about Joseph Campbell, comparative mythology, and more!

Moderators: Clemsy, Martin_Weyers, Cindy B.

Locked
tonyd
Associate
Posts: 51
Joined: Wed Apr 16, 2008 3:23 pm
Location: Dublin, Ireland
Contact:

An ontology for cultural history using Campbell's corpus

Post by tonyd »

There are different ways of looking at things depending on whether the viewpoint you choose is Christian, Marxist, Neo-Platonist etc. Some such viewpoints may become enshrined in how data is structured on the web in the near future due to the advent of linked-data, the shadow web of data that is gradually becoming more important.

Linked-data, and associated ontologies, is an involved topic and from a Campbellian point of view one worth playing around with for a few years. I've started an ontology that, using Campbell's corpus as a blueprint, tries to structure data pertaining to cultural history. The project is 'Cultural History' on Stanford's webprotege.

For a visualisation of the shadow web see: Linked Open Cloud (LOD) data: http://lod-cloud.net/versions/2014-08-30/lod-cloud.svg

For those interested to see how this sort of data is accessed go to: http://dbpedia.org/snorql/
and in the white box delete what's there and paste in:

Code: Select all

PREFIX dbo: http://dbpedia.org/ontology/
PREFIX yago: http://dbpedia.org/class/yago/
SELECT ?fairytale ?abstract
WHERE
 { ?fairytale a yago:Fairytale107222581 .
 ?fairytale dbo:abstract ?abstract .
 FILTER (regex(?fairytale, "king","i"))
 FILTER (! regex(?fairytale, "daughter","i"))
 FILTER(lang(?abstract)="en")

 }
ORDER BY ?fairytale
enclosing the two web addresses on the first two lines in '<' and '>' so '<' in front of http:// and '>' after ontology/ and yago/ as in the pre existing prefixes already listed above on the http://dbpedia.org/snorql/ page .

Basically this query looks for things that have as 'object' the yago fairy tale category and an associated abstract if it exists. It looks for 'king' in the title but discards tales with 'daughter' and, as it would return all language versions, we confine it to 'en', or english results. The '?' denotes a variable name and the 'WHERE' clause always consists of triples of subject-predicate-object type which are joined by '.' Try variations!

This sort of thing did form part of Tim Berners-Lee's original vision for the web. http://www.w3.org/DesignIssues/Architecture.html

There is much about the Linked Open Data cloud that is annoying; this is because it is designed to accommodate machines not people. One emerging hub is DBpedia. The information it pulls from different sources on Joseph Campbell (including Wikipedia) can be seen at: http://www.dbpedia.org/page/Joseph_Campbell. Most of the info looks fine but we note some items that are questionable and also a lack in certain areas. For instance that Campbell was author of the four volumes 'Masks of God'. For reference the quickest way to influence such information is to change the infobox on the Wikipedia Joseph Campbell page (DBpedia only wants facts so it ignores discussions or articles and strips the infobox information from every Wikipedia page). The monomyth continues to embed itself in the cultural psyche: http://dbpedia.org/page/Monomyth but Campbell's later work, in my view, is greatly under represented.

The different prefixes i.e dbo: are abbreviations for full web addresses to where each term is defined. So essentially each prefix refers to a different vocabulary, or set, of terms. Many of you will be familiar with Zotero it uses four such vocabularies and the Mellon supported Content Management System, Omeka, is underpinned by two such vocabularies.

WordPress has plugins for producing linked data.
Tony D.
Fare Forward

Locked