[Top] [All Lists]

Re: [ontolog-forum] Two ontologies that are inconsistent but both needed

To: Pat Hayes <phayes@xxxxxxx>
Cc: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>, Pierre Grenon <pierre.grenon@xxxxxxxxxxxxxxxxxxxxxx>
From: "Smith, Barry" <phismith@xxxxxxxxxxx>
Date: Mon, 11 Jun 2007 14:53:17 -0600
Message-id: <20070611210030.2FE8610950D@xxxxxxxxxxxxxxxxx>
At 10:16 PM 6/10/2007, Pat Hayes wrote:
>>At 05:39 PM 6/8/2007, Pat Hayes wrote:
>>>>An example is described here:
>>>Hey, nice survey. Utterly wrong in its conclusions, but nice :-). 
>>>I'm happy to welcome you to this debate which many of us have been 
>>>involved in for quite a long time. (see for example
>>>http://www.ihmc.us/users/phayes/Endurantism&PerdurantismDebate2002.pdf )
>>>But you come to the wrong conclusion. These two 'irreconcilable' 
>>>ontologies ARE reconcilable, if one does things right. The basic 
>>>error is to assume that what a philosopher means by 'exists' has 
>>>to be rendered into the logical existential quantifier. That is 
>>>good form, perhaps good doctrine, when the game is to use formal 
>>>logic to sharpen philosophical debate; but that is not (or at any 
>>>rate should not be) what we are trying to do here. The only 
>>>sensible engineering attitude to take towards the logical 
>>>existential quantifier is that it means "is an entity which can be 
>>>referred to", i.e. an entity which is the denotation of a logical 
>>>term; which as long as we are using a reasonably classical logic 
>>>is essentially vacuous, of course. In a pluralistic ontological 
>>>framework, this cannot usually be interpreted as any philosopher's 
>>>notion of existence. Those notions have to be treated as classes 
>>>or properties. Yes, existence IS a predicate, when there are many 
>>>notions of existence to be considered. It has to be in any logic 
>>>which is intended to support interoperability. (See the 
>>>regrettably brief discussion at 
>>>  . Sorry, I know that to say this to a philosopher is like farting in 
>>>After thinking and arguing about endurance and perdurance for 
>>>longer than I care to remember, I have come a rather mundane 
>>>conclusion which can be summed up as follows: the 
>>>continuant/occurrent distinction is basically a distinction 
>>>between *how we use names* when talking about spatiotemporal 
>>>entities. It should not be seen as a fundamental ontological 
>>>distinction: it is merely a linguistic distinction between modes 
>>>of expression. Things we call continuants are things for which we 
>>>tend to use the same name at different times, so it is natural to 
>>>encode changes to their properties by attaching the temporal 
>>>parameter to their properties and relations rather than to them: 
>>>we write things like
>>>(inside Fritz Bratwurst Morning)
>>>but we don't tend to talk of Fritz having temporal parts. Special 
>>>terminologies are used to distinguish these temporally-sensitive 
>>>relations and properties: "fluents", "roles".
>>>Occurrents, on the other hand, are things that we do tend to speak 
>>>of as having temporal parts or 'episodes', so it is natural to 
>>>formalize temporally-relative talk of those entities by attaching 
>>>the temporal qualifier to the name itself. If Fritz and the 
>>>Bratwurst were occurrents, we might write
>>>(inside (episode Morning Fritz)(episode Morning Bratwurst))
>>So there is a distinction between continuants and occurrents which 
>>is prior to our use of names -- for otherwise in virtue of what 
>>would we attach the first kind of name to the first kind of entity 
>>and the second kind of name to the second kind of entity?
>Simply from linguistic habit.    (01)

If you apply term A to some sorts of things, and term B to other 
sorts of things, then there has to be something about the As (the 
things which get called 'As') and the Bs (the things which get called 
'Bs'), which allows us to make the assignment. It does not seem to be 
entirely random.    (02)

Moreover, it seems to be a habit which we all share, and are good at 
exercising. Yet more evidence that there is some easily apprehendable 
corresponding difference on the side of the entities.    (03)

>My point is that such a habit does not necessarily indicate a true 
>ontological distinction between kinds of entity. (At most, it would 
>likely be a difference in scale or social importance. We give some 
>atmospheric rotary events names, not others, based largely on the 
>wind velocity in them.)    (04)

Scale would precisely be the sort of differences I have in mind. 
(Social importance wouldn't work, because that too would rest upon 
some underlying [though perhaps oft' changing] difference on the side 
of the entities.) But here it is not one or scale. What is it then?    (05)

>And in any case, for almost any such habit one can find cases where 
>the opposite convention is used, and yet which are widely understood 
>without explanation: which surely ought to be impossible or 
>incoherent if indeed it reflected a true ontological difference 
>between kinds of entity.    (06)

This sounds like a famous (bad) Wittgensteinian argument against p 
(for any 'p'). "There might be people behind a distant hill who believe not-p."    (07)

>>>If one puts all philosophical discussion aside for a moment and 
>>>asks for a purely formal, syntactic, way of distinguishing these 
>>>ways of describing things, what it seems to amount to is where to 
>>>attach a temporal parameter to a time-free assertion. One might 
>>>pose it as a challenge: given that
>>>(inside Fritz Bratwurst)
>>>is true during a time-interval
>>>invent a systematic way of encoding that fact by incorporating the 
>>>temporal parameter into the logical expression. There are 
>>>basically three places it can go: attached to the entire 
>>>expression (the 'ist' version:
>>>(ist Morning (that (inside Fritz Bratwurst)))
>>>using the paraphernalia of context logic), or attached to the 
>>>relation symbol (the first option) or attached to one or more of 
>>>the argument terms (the second option). These correspond 
>>>respectively to the hybrid/context-logical, continuant and 
>>>occurrent ways of treating time.
>>>So, can these co-exist? Yes, of course. One can use both (in fact, 
>>>all three) modes of expression in a single ontology, and in a 
>>>reasonably expressive logic (like IKL) can even write axioms which 
>>>relate them systematically. One does need to use some discipline, 
>>>to keep things straight. One has to use even more discipline to 
>>>use them both (or all) in ways that respect the philosophical 
>>>prejudices of all users. For example, if someone insists, as you 
>>>do, that it is incoherent or irrational to talk of temporal parts 
>>>of a continuant, then one will probably need some kind of 
>>>mechanical check to ensure that no entity is ever spoken of in 
>>>both temporal styles. Such code could be written, but I personally 
>>>see no practical use for it, and large amounts of harm caused by 
>>>insisting upon the distinction it would be there to check.
>>>The continuant/occurrent distinction seems to be of no actual 
>>>value in real ontology engineering[1]: on the contrary, in fact, 
>>>recent discussions on this very list and on 
>>>public-semweb-lifesci@xxxxxx seem to illustrate what I have always 
>>>found to be the case, that as soon as one gets away from nice 
>>>homely examples like Fritz' bratwurst, the distinction becomes 
>>>more and more tenuous, intuitions regarding it dissolve, and the 
>>>insistence on its being a basic distinction rapidly becomes more 
>>>trouble than it is worth, causing long and pointless debates and 
>>>tending, if anything, to produce new, artificial barriers to 
>>>interoperability rather than help with our practical goal. The 
>>>real world is full of entities which are both 'continuant' and 
>>>'occurrent', both thing and process: ocean waves, storms, weather 
>>>fronts, the Olympic flame, a cumulus cloud, the interior of a Bessemer 
>>all of these are self-evidently continuants, to me; some of them 
>>even get proper names ("Tropical Storm Barry", anyone?)
>I know they do. So do some tsunamis and earthquakes. But my point is 
>that they are also obviously process-like in nature. They don't fit 
>either category properly: or, perhaps better, they fit both of them 
>to some extent or for some purposes. So does it follow that there 
>were two Hurricane Ivans that demolished Pensacola at the same time, 
>the process (it is very hard not to see a hurricane as a process 
>when one has been inside it) and the continuant? Surely this is 
>ridiculous: just because we give something a name, it has to be 
>regarded as being a different *ontological* category?
>>>the Krebs cycle
>>is a series of enzyme-catalysed chemical reactions, so 
>>self-evidently an occurrent
>And yet it has a name.
>>>a tomato ripening on a sunny windowsill,
>>is a tomato, so no problem, eh?
>No, a very nasty problem. It is a thing *and* a process. Most things 
>are also processes (Whitehead).
>>>a cell expanding because the sodium pumps in its membrane are 
>>>insufficient to oppose the osmotic pressure,
>>is a cell, so ditto.
>You are making my point for me. Im sure you can give snappy answers 
>about which kind of thing you think these all are. But (1) others 
>might give equally snappy but different answers, but more seriously 
>(2) some of things one wants to be able to say about them are then 
>made illegal, which makes no ontological or pragmatic sense, and (3) 
>nothing useful follows from this distinction, which also suggests 
>bad engineering at work. I don't want to argue with you or anyone 
>else about whether or not a tomato is a continuant: I don't CARE if 
>is one or not. But I do want to be able to talk about it as a thing, 
>and also as a process, without being told I can't, for essentially 
>philosophical reasons, do that.    (08)

The principal biological ontologies make the distinction clearly, and 
can talk perfectly well about continuants where needed and about 
occurrents where needed. They are also increasingly well engineered.    (09)

>>>Ive never seen any convincing pragmatic or engineering argument 
>>>for insisting on this as a rigid distinction.
>>    (010)

One pragmatic argument is that it would be at this stage practically 
impossible to turn great ships of state such as the gene ontology 
through the necessary 180 degrees in mid-flow.    (011)

BS    (012)

>>I used to be a philosopher. Now I am an ontologist.
>I have always been an ontologist :-)
>>The difference is that now I do not engage in those philosophical 
>>discussions for which it is clear that there are good arguments on either 
>I was not engaging in a philosophical discussion. My case was, 
>speaking purely as an ontologist, that there are no good pragmatic 
>ontological arguments for insisting upon the sharp distinction. IMO, 
>to the extent that they do insist upon it, both DOLCE and BFO are 
>fundamentally flawed.
>>  I just choose (like driving on the left). I choose to distinguish 
>> between continuants and occurrents.
>Fine. But then the question arises as to whether your ontological 
>framework, which requires this distinction, is of more use than a 
>similar one which does not.    (013)

I think we need to test this empirically. So far we are winning -- 
the GO is, by several measures,the world's most successful ontology.    (014)

>I would also observe, before leaving this topic, that your followers 
>often do not take this casual I-chose-the-left attitude to this 
>distinction. They often, indeed, insist that this distinction MUST 
>be made, that any ontological framework which does not make it is 
>inherently wrong, and cite you as an authority. (This has happened 
>three times when I have been present in the last year, most recently 
>at an NSF meeting.)    (015)

Perhaps they are just scared of driving on the right in Middlesborough.    (016)

>And I have heard you make the case vehemently, in public forums of 
>one kind or another. The paper you cited, to be published in the 
>future, makes this case in very strong terms, and indeed seems to be 
>devoted to little else.    (017)

Perhaps I am just scared of driving on the right in Middlesborough.    (018)

>Should I take it that my impression from all the above is mistaken, 
>and that you do not believe that this distinction is inherently 
>important, but that it amounts simply to a convention?    (019)

You are trying to tempt me once again into engaging in a 
philosophical argument.
BS     (020)

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    (021)

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