[Top] [All Lists]

Re: [ontolog-forum] (OT) German

To: "[ontolog-forum] " <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Pat Hayes <phayes@xxxxxxx>
Date: Fri, 11 Jan 2008 08:57:35 -0800
Message-id: <p06230903c3ad4cf0be54@[]>
At 4:03 PM -0500 1/8/08, Stavros Macrakis wrote:
>On Jan 8, 2008 2:34 PM, Duane Nickull 
>It was meant to be funny however a hippo is 
>still quite a different beast from the horse.
>Let me try to save this discussion from utter triviality.
>Lakoff has pointed out    (01)

And many others have disagreed with him. Just a warning note.    (02)

>  the centrality of metaphor, analogy, and graded 
>categories in natural language, contrasting them 
>with the property-based strict categories of 
>Aristotle.  The female biological and genetic 
>parent who lives with and cares for a child, is 
>its legal parent, and is married to and lives 
>with and its biological and genetic father is 
>certainly a "mother" in every traditional way, 
>but how do we deal with cases like adoptive 
>mothers (cares for but not biological), foster 
>mothers (cares for but not legal parent), 
>lesbian co-mothers (one of whom may or may not 
>be a genetic or biological parent), surrogate 
>mothers (biological but not genetic), egg donors 
>(why don't we use the word "mother" here?), 
>stepmothers (married to father), birth mothers 
>(biological but doesn't care for or live with), 
>transgendered adoptive mother (not genetically 
>female), etc.? The intersection of the 
>properties of these different kinds of "mother" 
>is empty.    (03)

Excellent example of why it is better, when 
trying to do ontology, to think about what is 
being decribed, ie what there is, ie do ONTOLOGY, 
rather than linguistics or conceptual analysis. 
That is, to describe the relationships in the 
world, rather than try to conform to some 
structure inferred from the use of words in the 
surface of NL. Your account above gives very 
strong hints as to what the relevant 
relationships here are: being a biological parent 
of, being a guardian of, being a legal parent of, 
being female, playing a female role, being 
married to, etc.. To cover all the possibilities 
you describe evidently requires some care in 
choosing these and some more care in describing 
their relationships: an adequate ontology for 
'mother' in all its complexity will be quite 
intricate. Far too intricate to expect that one 
will be able to take an ontology for 'normal 
mother' and simply apply an ontology modifier to 
it to get the right ontology for, say, adoptive 
mother. (Or at any rate, the description of such 
an ontology modifier would be best tackled after 
the basic ontology work is already done.)    (04)

>How to treat these in an ontology?
>-- with exceptions?  "A foster mother is a mother except that ...."    (05)

I doubt this will work, though Cyc have claimed 
some success with 'microtheory mappings' of this 
kind. The mappings have to be able to perform 
quite complex operations of the ontologies, 
though: simple default-overriders are not 
adequate.    (06)

>-- with limited analogy? "A foster mother is 
>like a generic mother in that X and Y are true"    (07)

That sound like part of an ontology of 
'likeness'. Why do you need this? If you do, then 
using analogy/metaphoric mappings (such as those 
found by 'structure mapping') is likely a better 
bet than  trying to incorporate it into the basic 
ontology.    (08)

>-- as arbitrary names for otherwise arbitrary 
>categories? "A foster-mother is like a mother 
>because she happens to share some properties 
>with a mother (just as a father shares some 
>properties with a mother), but also shares the 
>substring 'mother' in the category name."    (09)

Again, why do you need this?    (010)

>-- with graded membership? "A foster mother is 
>50% mother." (not clear what sort of reasoning 
>you can do on this...)    (011)

Blech. Fuzzy logic, anyone?    (012)

>And how do you characterize the modifier "foster" or "step" or "adoptive"?    (013)

You don't, see above. The purpose is not to 
account for the complexities of English, remember.    (014)

Pat    (015)

>             -s
>Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/ 
>Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/ 
>Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
>Shared Files: http://ontolog.cim3.net/file/
>Community Wiki: http://ontolog.cim3.net/wiki/
>To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>    (016)

IHMC            (850)434 8903 or (650)494 3973   home
40 South Alcaniz St.    (850)202 4416   office
Pensacola                       (850)202 4440   fax
FL 32502                        (850)291 0667    cell
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes    (017)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (018)

<Prev in Thread] Current Thread [Next in Thread>