[Top] [All Lists]

Re: [ontolog-forum] Ontology similarity and accurate communication

To: "[ontolog-forum] " <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Pat Hayes <phayes@xxxxxxx>
Date: Wed, 12 Mar 2008 10:25:09 -0500
Message-id: <p06230901c3fd9f97cb96@[]>
At 11:27 PM -0400 3/11/08, Patrick Cassidy wrote:
Content-Type: multipart/alternative;
Content-Language: en-us
Responses to selected comments:
[PC] >> A definition is a description of the meaning of a term, whether it is a linguistic definition or a logical specification. That is the analogy.
[PH] > Not good enough. First, tell me what you mean by 'meaning of a term' when we are talking about formal ontology languages.
  I use the term in this sense: In a formal ontology there is the logical meaning, which consists (for an instance of a type) of all of the inferences that can be drawn from the fact of something being an instance.

OK, thanks for that definition of 'meaning'. But this, if read strictly, is a very weak notion. Very little can be inferred from the fact of being a class instance alone. One usually needs more axioms about such instances in order to draw any nontrivial conclusions.

[PH] > The burden of proof is on you. Cite a reference explaining how we all come to have identical mental models.
    Quite the contrary. I have merely asserted that I have a hypothesis, and presented the evidence that causes me to form that hypothesis.

What evidence? You yourself have agreed that human agreement on word meanings does not imply ontological agreement, yet this was the whole basis of your hypothesis, as far as I can tell.

  Any hypothesis needs to get extensive evidence before it will be believed by anyone other than the hypothesizer.  That is well understood.  Forming a hypothesis, even from vague analogies, is one of the methods for developing a motivation to do research, and I have presented it as nothing more than that.

Well, no, you have claimed to have an argument for it, and much of the debate so far has concerend the merits (or otherwise) of that argument.

 You don¹t think the evidence I cite is in any way convincing, and doesn¹t justify research.  Fine, but:
      You, on the other hand, have asserted in very strong terms different variations on the theme that there is strong evidence against my hypothesis

No, against your argument for it. And there is some - admittedly anecdotal, but one can perform a similar experiment anywhere and will quite quickly find similar results (it happens regularly in knowledge elicitation experiments, for example, which IHMC does a lot of) - evidence against it, which I find compelling. As the proposal (the 99.9% correspondence) is a very strong claim, it requires only a little evidence to refute it.

, and yet the only evidence you have provided is that some adult ontologists couldn¹t agree on some representations in some cases ­ which is only remotely relevant, because it did not involve an attempt to agree on the concepts corresponding to the terms in the basic vocabulary.

As you have never told us what is in is vocabulary, I am unable to respond; but 'in' and 'room' seem like pretty simple concepts to me.

   If you say something is disproven, the burden is on *you* to cite the published evidence.  But you haven¹t cited a single published source, let alone one that comes close to being on point.  If you don¹t want to dig for the references you are sure must be there somewhere, I sympathize, but then don¹t keep saying that the hypothesis of a strongly similar mental model for basic terms has been disproven.

I say there is evidence against it. I also say that as a hypothesis it raises enormous problems immediately, such as: how could one possibly verify it; how could it occur in nature; etc. . None of your responses to these points have been even logically adequate, and you do not appear to be aware of the difficulties.

 I accept your disagreement.  If you want to assert that you have better evidence against the hypothesis than I have for it, please present that evidence.
[[[3]]]  [PC] >> Clearly, if some entity is an instance of one ?occurrent¹ but not an instance of another ?occurrent¹,
[PH > That does not make sense. Occurrents do not have instances.
    You misinterpreted.  The word occurrent was quoted to indicate the type.

Ah, I see. The quote marks seem to have got scrambled by some lexical garbling in the first message. OK.

 To clarify, a restatement:
     Clearly, if some entity is an instance of the type ³occurrent² in one ontology, but not an instance of the type ³occurrent² in another ontology, then the meanings of ³occurrent² in the two ontologies must differ because they are using the terms in different senses.
[PH] > This whole process [getting agreement] is a fantasy. This simply does not happen. This is like assuming that if you get two people together and they do enough careful exegesis of one another's beliefs, they will eventually agree that they have the same religion.
     Now the basis for your skepticism is becoming clearer.  You apparently think that most ontologists¹ ideas on how to formalize domains are merely beliefs that are insusceptible to change by argumentation.

I wouldn't say 'merely', but yes indeed, these positions are most stubbornly held and are indeed resistant to argumentation. (Have you ever tried to argue with a philosopher?)

 It¹s true, I don¹t have such a low opinion of most ontologists.

I don't think of this as a low view. On the contrary, I kind of admire people who stubbornly hold to their positions and defend them against attacks with good arguments. Where we differ is that I allow that there can be genuinely different points of view, while you apparently don't.

  We differ there.  Of course, that may well be true of some ontologists.  Do you think that  any of them will admit it?

Of course they admit it. They rejoice in it; they write books and papers about it.

  [PC] >> We would need to look at the specifics of the representations to determine what the problem is.   Perhaps not something that can be done in an on-line discussion, but maybe worth trying.
  [PH]  There are entire libraries written on this division of perspective. It has been done to death on various archived ontology forums, also. But when you say this is a 'problem', what do you mean, exactly?
   I am talking about trying to resolve any differences one by one, by trying to understand the fundamental components of each view that causes the logical contradiction.

So am I.
 For 4D/3D, determining whether there is a consistent ontology encompassing both views depends on the specific formalizations that are inconsistent.  It¹s easy to create a contradiction.  I say P, you say not-P.  The question I would ask is whether the motivation for the different representations is based on different requirements

That certainly seems to be the case, yes.
and whether it has to be in the foundation ontology.

Well, again I have no idea what you mean by this (and don't believe that such a thing exists) but it is hard to see how any 'foundation' can avoid taking some position on how to describe time and change.

 Once again I will assert the opinion (no proof needed) that the likelihood of getting agreement among a group depends on the motivation.  Just firing off notes like these is approximately zero motivation.  A serious project to develop a foundation ontology corresponding to the linguistic defining vocabulary might provide more motivation.  Not for everyone, but we don¹t need everyone, just a large enough core to sustain serious collaborative effort.

This particular debate has been raging on and off, not just in random email notes, for at least 20 years.

  [PC] >> Question: do you think that equating a zero-time-interval timeslice of a person with a 3D ?Continuant¹ person would lead to any logical incompatibility?
   [PH] > I KNOW that it would. Here's a quick one:
>  (cl-comment 'basic continuant axiom'
>    (forall ((t Time)(c Continuant))
>               (= (c at t) c ) ))
>  (cl-comment 'definition of age property in 4-d'
>      (forall (x (t Time))
>         (=  (age (x at t))
>               (minus t (birthtime x)) )))
>  from which, and a little arithmetic, it follows that
>      (forall
>           ((t Time)  (s Time))
>          (= s t) )
>  i.e. time is impossible.
    I¹m not sure how to interpret this argument.

I did not expect that you would be.
The set of axioms conflates diachronic identity (the = in the continuant axiom) with mathematical identity (the = in the 4-d axiom) although the two are very different concepts.  I know you are well aware of that

The only kind of identity I am concerned here with is logical identity, which means being the very same thing: (= a b) is true in I just when I[a] is the same element of the universe as I[b]. This carefully avoids any commitment as to what 'kind' of thing this single thing with two names is.

, so why are they there together, looking as though they are the same thing?

I am using the equality sign in the same way, with the same meaning, throughout.

Is this your way of saying that the ?continuant¹ view is incoherent because it conflates diachronic and mathematical identity?

No, I don't think it is incoherent. It is simply inconsistent with the other way of looking at things, is all.

 I like to use the 3D formalism for most of my representations, but would never use the ?=¹ sign like that.  If I felt I needed such an axiom, I would define a relation ?isDiachronicallyIdenticalTo¹ to express that notion.

And what properties does this have? (Does it satisfy the Leibnitz condition? Is it an equivalence relation?) How is it related to logical equality? (If a=b and b diachrinicallyidenticalto c, does a=c?) How would you express the continuantist dictum that a continuant is identically the same thing at all times it exists?

  Who uses such an axiom?  A Google search for ³basic continuant axiom² found no hits.

It is my best effort to state the continuant dictum in a 4-d linguistic framework of temporal parts. Admittedly this is not how a continuantist would likely express it: they would be more likely to declare that expressions of the form (c at t)  -  which, by the way, should be written (at c t) to be legal CLIF - are syntactically illegal, because meaningless. This would eliminate such immediate contradictions as the above, but it will instead produce a plethora of syntax errors when the two ontologies are mixed together.


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

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    (01)

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