uom-ontology-std
[Top] [All Lists]

Re: [uom-ontology-std] uom-ontology-std - strawman UML

To: ingvar_johansson <ingvar.johansson@xxxxxxxxxxxxx>
Cc: uom-ontology-std <uom-ontology-std@xxxxxxxxxxxxxxxx>
From: Ed Barkmeyer <edbark@xxxxxxxx>
Date: Fri, 07 Aug 2009 12:28:40 -0400
Message-id: <4A7C5638.3050706@xxxxxxxx>
ingvar_johansson wrote:
> I agree that a fourfold distinction as yours below between Q1-Q4 is
> necessary. However, to me you are using 'category' in an unusual sense,
> and I think that in order to avoid confusion one should distinguish
> between two senses of 'being an instance of', and then use different terms
> for them. On the one hand we have 'an instance' as a spatiotemporally
> located individual entity, this I will call 'an instance'; on the other
> hand we have 'an instance' as a determinate of a determinable (red is an
> instance of color, 14 kg is an instance of mass, etc.), this I will call
> 'a determinate'. In my opinion, VIM confuses these two senses.
>
>   
I didn't realize that I was using "category" in an unusual sense.  I 
used the term to express the inverse of the subsumption relationship:  A 
'is a category of ' B iff B subsumes A.  And in that I meant to imply 
that A is a unary relation (a class, the sense of "category" that I 
assume is "usual") in its own right, i.e. that A(x) is true if the 
individual x has the characteristics that define A and false otherwise.  
I would be happy to change all occurrences of 'is a category of' to 'is 
a subtype of'.  I am less comfortable using 'is a kind of', because I 
think it has an ambiguous model.     (01)

I never use "'an instance' as a spatiotemporally located individual 
entity".  I prefer the term "individual" for things that appear in the 
domain of discourse (which is a potentially larger population than 
"spatiotemporally located individual entity"), in order to avoid 
precisely the confusion you identify.  Further, I hope I only used 
"instance" in the relation "is an instance of", with the 
interpretation:  x 'is an instance of' A iff (x is an individual in the 
domain of discourse) and A(x) is true.  You may use "determinate of a 
determinable" for that relation if you please, but I doubt that it will 
be easily comprehended by much of our audience.    (02)

> Furthermore, the fourfold distinction in question applies even to
> non-quantified properties, and I think that to bring this out can help to
> clarify the issues at hand. Let me exemplify with the determinable 'shape'
> and one of its determinates, 'spherical'. A necessary requirement for a
> quantification is that all the determinates of certain determinable can be
> and are linearly ordered, and this has not yet been done for the shape
> determinable.
>
> Q1-spherical = the spherical instance of ball A at t, the spherical
> instance of the planet B at t, etc. That is, Q1 = determinate property
> instances.
>
> Q2-shape = the shape instance of ball A, the shape instance of the planet
> B, etc. That is, Q2 = determinABLE property instances.
>
>   
The examples are confusing, but the definition "Q2 = determinABLE 
property instances" is what I meant.  I think the example of Q2-shape is 
"spherical" or "sphere", as distinct from "prismatic solid".  So "ball A 
at t" is a 'determinate' that 'is an instance of' the 'determinable' 
"sphere".  Thus the 'individual' "sphere" 'is a subtype of' Q1, and the 
same 'individual' "sphere" 'is an instance of' Q2.
> Q3-spherical = the determinate universal 'spherical' if you are a realist,
> and the equivalence class of 'spherical property instances' if you are a
> resemblance nominalist. That is, Q3 = determinate universal or determinate
> equivalence class.
>
> Q4-shape = the determinable universal 'shape' if you are a realist, and
> the equivalence class of 'shape property instances' if you are a
> resemblance nominalist. That is, Q4 = determinABLE universal or
> determinABLE equivalence class.
>
> And here are some axioms:
>
> A1: Necessarily, if a Q1-instance (e.g., a spherical instance) then also a
> Q2-instance (shape instance).
>
>   
No.  This is not an axiom I mean.  What I want is:    (03)

(forall (x D)(if (and (Q2 D) (D x))
                      (Q1 x)
                  ))
Every instance of Q2 is a subtype of Q1.
> A2: Necessarily, if a Q2-instance (e.g., a shape instance) then also an
> instance of a determinate of Q2 (instance that is spherical, or square, or
> star-formed, etc.).
>   
Again.  I don't understand this axiom.  What I mean is:    (04)

(forall (x)(if  (Q1 x)
                   (exists (D) (and (Q2 D) (D x)))
              ))
Every individual that is an instance of Q1 is also an instance of some 
determinable D that is an instance of Q2.
> A3: Only quantified determinates of the same determinable can be added and
> subtraced (in a physically menaingful way).
>
>   
The VIM term is "comparable" ("added and subtracted" is perhaps too 
specialized a notion).
But the axiom is stronger than that:    (05)

(forall (x y)(iff (exists (D) (and (Q2 D) (D x) (D y)))
                      (comparable x y) ))
x and y are comparable if and only if there is some determinable D in Q2 
such that x and y are both instances of D.    (06)

> A4 (for many but not all determinables): Two Q2-instances cannot possibly
> exist in the same spatiotemporal region.
>   
I am going to assume I don't understand this, because what I understand 
it might be intended to mean is false.    (07)

Assuming the relation (st-location x r) means x has a spatiotemporal 
existence in region r, then we probably do mean (although I am not sure):    (08)

(forall (x)(if (Q1 x)
                  (exists (r) (st-location x r)) ))
Every instance of Q1 has some spatiotemporal location.  Every measurable 
instance is physical.    (09)

What I think Ingvar means is:    (010)

(forall (x y rx ry)(if (and (exists D)(and (Q2 D) (D x) (D y)))
                                    (st-location x rx)
                                    (st-location y ry))
                             (not (= rx ry)) ))    (011)

If two 'determinates' (measurable instances) x and y are both instances 
of a common 'determinable', their spatiotemporal locations cannot be the 
same.    (012)

But I don't think that is true, at least not of what I called Q2.  For 
example, the width of a given box and the height of the same box are 
both instances of a Q2 (determinable) called "length".  Do we really 
mean that those properties are not coincident in space-time?  That seems 
to me to be an issue in philosphical geometry.  ;-)    (013)

-Ed    (014)

-- 
Edward J. Barkmeyer                        Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8263                Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263                FAX: +1 301-975-4694    (015)

"The opinions expressed above do not reflect consensus of NIST, 
 and have not been reviewed by any Government authority."    (016)


_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uom-ontology-std/  
Subscribe: mailto:uom-ontology-std-join@xxxxxxxxxxxxxxxx 
Config/Unsubscribe: http://ontolog.cim3.net/mailman/listinfo/uom-ontology-std/  
Shared Files: http://ontolog.cim3.net/file/work/UoM/  
Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM_Ontology_Standard    (017)

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