What is the difference between a Taxonomy and an Ontology?

OWL Knowledge

By Rick Jelliffe
February 7, 2010 | Comments: 4

Q. What is the difference between a Taxonomy and an Ontology?

A. Not much.

The wisdom of the masses has the following:

A taxonomy might also be a simple organization of kinds of things into groups, or even an alphabetical list. However, the term vocabulary is more appropriate for such a list. In current usage within Knowledge Management, taxonomies are considered narrower than ontologies since ontologies apply a larger variety of relation types.

So we have a hierarchy ...err... spectrum:


Vocabulary
A simple list or grouping of kinds of things, made by muddlers

Taxonomy
Err something in between, probably arranged as a hierarchy, made by plodders

Ontology
Any kind of relations between kinds of things, made to disconcert plodders and muddlers

Lets see how XBRL uses the terms, for example:

Taxonomies ... are the reporting-area specific hierarchical dictionaries used by the XBRL community. They define the specific tags that are used for individual items of data (such as "net profit"), their attributes and their interrelationships. Different taxonomies will be required for different business reporting purposes.

So you have tags (vocabularies) but they are arranged hierarchically, so they are taxonomies? But are they ontologies?

Well, presumably not, since XBRL has a (semi-official?) effort to make XBRL Ontologies. And they are called Ontologies because they use the W3C OWL. The subject matter or XBRL Ontologies is expressing citations and bibliographic references in RDF.

Now OWL says is OWL can be used to explicitly represent the meaning of terms in vocabularies and the relationships between those terms. This representation of terms and their interrelationships is called an ontology.

The use of ontology is a little problematic, because it seems to me that it is one of those terms which has ended up meaning in one discipline (computer science) exactly the opposite of what it meant in its original discipline (philosophy).

In information science, an Ontology is a set of types, properties, and relationship types about some knowledge domain; in philosophy it seems to me to be one step more meta: it is concerned with categories of being (though not Bill Clinton's questions on what 'is' means, probably) and not facts obtained about (beings) or particular properties relating to them: it is about what exists and what kinds of existences we can talk about.

(This confused me a lot, a few years ago: at first I thought these information guys were really ignorant, then when I found everyone used ontology in this loose way, I decided I was really ignorant: my conclusion now is that the terms have diverged to opposite meanings and the ignorance is to expect one discipline's terminology to be exactly the same as another's. What good is pedantry?)

I can see that Ontology is a useful word: a bigger, smarter, more abstract brother of Taxonomy. But it would be nice if there were some term to replace Ontology in the classical philosophical sense (if I get it right, IANAP) meaning a meta-ontology.


You might also be interested in:

4 Comments

Yes, ontologies have a larger variety of relation types. Yes, you can choose to participate in debates that compare and contrast the computer science notion of ontology with the notion owned by philosophists. However, contextually, given you put the question to the O'Reilly membership, the computer science notion rules:-} If you still want diversity , each version of OWL expects a differing set of properties as rules.

Carl: :-)

I came from philosophy into computer science and found the shift hard, but after a while no big deal. I use "ontology" all the time in the CS way. Indeed, I tend toward more and more restricted definitions: I'm happy to leave out the conceptual stuff and talking about specific classes of logical theories. Indeed, I tend to distinguish taxonomy from ontology on the bases of logical sophistication, e.g., in a taxonomy all subsumptions are derivable from asserted paths whereas in an ontology, part of the taxonomic structure is implicit.

Even within CS, you can get people moaning about how people have corrupted that grand old term, "Ontology" :)

YMMV.

In order to meet the SEC filing needs Xteron has come up with an Online XBRL compliance solution which addresses all the SEC requirements in the most efficient manner. Our product with simple easy to use workflows is a low cost solution without compromising on Quality and is becoming highly popular among the XBRL user community.

For more information please visit www.xteron.com or email support@xteron.com

News Topics

Recommended for You

Got a Question?