Cognitive Dissidents

I've recently read the article about by Clay Shirky about how Ontology is Overrated.

That's how long I was offline and how long it has taken me to catch up - luckily while I was catching up I made a lot of notes on my internal blog - the one I use to write reminders and tech notes to myself - and now some of these are going to get refined and see the light of day.

My major problem with the Clay's argument is that it conflates artificial categorisation taxonomies - in this example the Library of Congress categorisation scheme - with naturally evolved taxonomies.

Human taxononmy is not machine taxonomy.

Human taxonomy is not a formal taxonomy - human taxonomy changes with new entries, new ideas and new concepts.

Human taxonomy is a folksonomy.

I have to say - from the point of view of a human being - "Ontology is Everything".

Consider the following oppostional "semantic features" - or "tags" as they are called when used in the "semantic web".

RAW vs COOKED
DEAD vs ALIVE
HOT vs COLD
FOOD vs POISON

This naturally occuring ontology is not just an "imaginary" structure created to organise objects - it is a vital method of "carving the world at its joints" that is necessary for human survival.

We all use this ontology daily in order to determine what is "safe to eat" - a category that is both culturally determined and culturally flexible.

I would have never have known that "RAW FISH" was safe to eat - and very nice too - if the Japanese hadn't overlaid their own ontology onto mine.

Because they placed the "RAW FISH" sub-category into the "FOOD" category - I modified my category - and now eat Sushi when I can.

Before that I would have placed "RAW FISH" in the "POISON" category - my mother still does.

Back when our ancestors created ontologies it was nothing more than a primitive survival mechanism.

Ontology is necessary - human beings use ontologies all the time to navigate the world safely - so for our survival "ontology is everything".

But I think Clay Shirky was building a "straw man" argument to make his point.

There are many problems with human created ontologies - "Taxonomies" - but I think that most of these are caused by using out-of-date taxonomies.

We need dynamically created taxonomies - one example would be the taxonomic tree used by biologists.

Every so often a new animal, fish or insect is discovered that doesn't fit with the existing taxonomy.

Biologists don't go - "oh no - it doesn't fit - we must junk our existing classification scheme".

Instead they debate long and hard about what generally agreed changes can be made to the existing taxonomy to make it reflect the new discovery - while trying to preserve as much of the existing taxonomic structure as possible.

So when Clay Shirky says that "Ontology is Overrated" - I think what he really means is that "static ontologies such as fixed taxonomies based on tree-like hierarchies are overrated" - a slight difference.

Any categorisation system that is flexible and dynamic will reflect the flexible and dynamic nature of human categorisation - and that will be more usable to human beings than a static ontology - solely because it reflects the dynamic nature of human social categorisation systems.

I believe that the making of the "semantic web" means making the machine web easier to use by humans - by tapping into the shared collective knowledge of human beings - not by using fixed machine ontologies or systematic SEO keyword search.

Why not?

Let's forget formal ontologies - lets "tag" - and when we are not tagging we shoud argue about the value of tagging, of folksonomies and ontology.

That's how new ontologies are created.

Tags: tags tagging folksonomy tagsonomy ontology semantics human semantics machine semantics tag clusters tag clouds tagcloud tagclouds categorisation syndicspace