Re: [ontolog-forum] Asserting subclasses of open ranges or domains

Bernard and Pat,

BV> So I will define "WashingMachine", "HomeCinema", "BathTub" and
 > whatever you like as subclasses of "DomesticAppliance".

That statement is useful to give some examples, but it is not
a definition.

BV> In short I have a list of sufficient conditions for this class,
 > but no proper necessary conditions, beyond the tautological fact
 > of being the value of "hasAppliance". Sure enough, if the ontology
 > had been set a century ago, new subclasses would have emerge since,
 > and an ontology set today does not want to preclude whatever the
 > 21st century will bring about in this domain.

I agree with Pat on this issue:

PH> The only thing that will do that is to somehow restrict
 > the range with necessary conditions.

The type label DomesticAppliance is an example of what I call
a 'role type'.  It specifies the role of something without
specifying any observable properties that characterize it.
Examples include Employee, Author, Pet, Weed, or Tool.

The alternative is a natural type, which is characterized by
properties that can be observed in any instance of that type.
Examples include HumanBeing, Dog, or Dandelion.

All employees are human beings, all pets are animals (except
for metaphorical extensions to pet rocks), and all weeds are
plants.  But there is no observable property that determines
whether a particular human being is an employee, a particular
animal is a pet, or a particular plant is a weed.

If you want to state a decent definition, I would strongly
recommend looking up the word or phrase in a good dictionary.
Dictionaries designed for human readability are not as formal
as a statement in logic.  But for most purposes, I would trust
a well-edited dictionary to be far more reliable than most
so-called formal ontologies I've seen.

Following is definition 2b in the Merriam-Webster 9th Collegiate
Dictionary (which I like because it gives the first known year
when a word appeared in some document):

    2b: an instrument or device designed for a particular use;
    specifically: a household or office device (as a stove, fan,
    or refrigerator) operated by gas or electric current.

Note that 'appliance' is defined as a subtype of two other
role types:  'instrument' or 'device'.  If you follow up those
terms and delete the word senses that are not relevant, you
eventually reach something like "a physical object used for
some purpose."

By the way, a good printed dictionary is still vastly better
as an aid to the conceptual analysis necessary for ontology
than those instant definitions you can get from the WWW.

John Sowa

Received on Wednesday, 9 September 2009 16:37:37 UTC