W3C home > Mailing lists > Public > www-tag@w3.org > March 2008

Re: URI Declarations [Usage scenario 1b]

From: Pat Hayes <phayes@ihmc.us>
Date: Mon, 3 Mar 2008 12:37:04 -0600
Message-Id: <p06230908c3f1eda399d6@[10.100.0.20]>
To: John Cowan <cowan@ccil.org>
Cc: John Cowan <cowan@ccil.org>, "Booth, David (HP Software - Boston)" <dbooth@hp.com>, "www-tag@w3.org" <www-tag@w3.org>
At 12:48 PM -0500 3/3/08, John Cowan wrote:
>Pat Hayes scripsit:
>
>>  And one can give many others, of course. But you are talking about
>>  essential properties - what in OntoClean are called, if memory
>>  serves, "rigid" properties. Whereas what David requires us to put in
>>  our declarations are "identifying" properties, those which are
>>  supposed to pin down a referent uniquely. These are not the same
>>  notion. Being human, as a property of a human, is necessary (rigid)
>>  but it is most emphatically not 'identifying' (until we get the point
>>  of there being only one of us around.) In fact, necessary properties
>>  tend to be of the least use for identification, precisely because,
>>  being necessary, they tend to be shared by everything else in the
>>  relevant category. If you want to identify me uniquely, you will do a
>>  lot better knowing my SS number (not essential or rigid in any way)
>>  than the essential fact that I am human.
>
>It's true that lots of essential properties aren't identifying in
>themselves, but the more essential properties you know, the better a
>job you can do at identification.

(a) I see no reason why essential, as opposed to contingent, 
properties are important here, and (b) no matter how many you have, 
you can never get a single referent pinned down.

>  Because your SSN is not essential,
>there may be another person who has the same SSN (such glitches are not
>supposed to happen, but do).

Or, more to the point, there may be another person who COULD have had my SSN.

>  Your DNA dump, though, *is* essential

Why? Seems to me I could have had different DNA and still have been 
me. Its too late to change it now, of course.

>,
>and (given that you are not an identical twin) completely identifying.

Indeed, it is identifying.

>
>We sometimes find it useful to treat non-essential properties like
>name+DOB or SSN as identifying, but such treatments are always heuristic
>in nature, always subject to being debunked thus:  "No, I meant quite
>a different James Earl Carter born on 1924-10-01."

All 'identifying' descriptions are subject to correction like this.

>
>This can come up in real life.  I once volunteered to help check
>election records to identify a possible election fraud, and thought I had
>found a hit -- but it turned out to be two sisters (or perhaps cousins)
>living in the same house with identical first and last names (presumably
>different middle names) and nearly indistinguishable signatures.  A close
>inspection showed that they had been signing each other's election
>cards at random for years, but there was no actual fraud involved,
>just a nearly inextricable confusion of identity.

Nice example. I wonder if they ever got confused, themselves?

>
>>  Overall, I'd suggest that this entire discussion is so hopelessly
>>  muddled that the proposal is best dropped before we generate even
>>  more confusion. At least, lets see some convincing examples of what
>>  might constitute an appropriate 'declaration' before proceeding. I
>>  don't think that "identifying properties" exist. Can anyone prove me
>>  wrong?
>
>As Quine says, there is no entity without identity, so if we wish to
>drop identification, we will have a very difficult time distinguishing
>things at all.

Quine is referring to identity, not identification. All he meant was 
that you have to be able to make sense of equations: does A=B or not? 
Identification in the TAG sense is something altogether more than 
this: it means something social - having a kind of conceptual 
location in some social space of objects which can be "identified" - 
and it means being able to get from the 'identifier' to the thing it 
'identifies'. None of this is in Quine.

>  If nothing else, it is identifying to say that object
>o of sort s occupied space-time coordinates (x, y, z, t) relative to
>some convenient origin.  (I introduce "sort s" because although my liver
>occupies some location at the present moment, I occupy it too.)

You had better have a powerful (and universally agreed) set of 
'sortal' descriptions available for all contingencies. Good luck with 
that project. (Was that liver-at-t a temporal slice of a 4-d 
liver-history? Or was it a liver-continuant, mentioned at a time?) 
And then you have to have some other way to handle the 
non-spatiotemporal entities such as abstractions of various kinds, 
Platonic things like numbers and symbols, psychological properties 
(where did I put that mood I was in?) and imaginary and fictional 
things. And in fact, I bet this wouldn't be considered a proper 
identifying description in David's sense, in practice. It would be 
pretty useless in a declaration.

>
>Quantum theory aside, we do not want to treat classical objects as if
>they were cloudlets in a buttermilk sky, where it makes no sense to ask
>"Is this one the same as that one?"  Too much of logic and science goes
>down the rabbit hole.

I agree. Identity is a basic assumption underlying all of semantics. 
RDF and OWL already have identity built into their model theories, so 
we don't need anything new to have that. Tag:identification is 
something else again, however.

Pat



>
>--
>Evolutionary psychology is the theory           John Cowan
>that men are nothing but horn-dogs,             http://www.ccil.org/~cowan
>and that women only want them for their money.  cowan@ccil.org
>         --Susan McCarthy (adapted)


-- 
---------------------------------------------------------------------
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
http://www.ihmc.us/users/phayes      phayesAT-SIGNihmc.us
http://www.flickr.com/pathayes/collections
Received on Monday, 3 March 2008 18:37:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:53 GMT