Re: [ontolog-forum] Just What Is an Ontology, Anyway?

["Godfrey Rust" <godfrey.rust@xxxxxxxxxxxxx>]

This is one of those threads where I find myself agreeing with both "sides" of the debate, because one (Bill) is expressing a problem that in my experience is unavoidable and the other (Matthew) is expressing one good way of minimising that problem (in this case, agreeing on standard terms/ontologies/schemas).

 

I believe that standards and mappings both have an essential and permanent place in our increasingly overlapping worlds. Some of our work is on standards for specific communities, and on mappings between them, and increasingly the boundaries change and perceptions of what constitutes a community or domain are thrown into question. A case in point: we work with the development and management of the DDEX ("music" industry) and ONIX ("book" industry) message standards, and also on vocabulary mapping between them (through a new initiative called the Vocabulary Mapping Framework - VMF). Just yesterday there was discussion on the ONIX forum about how to use ONIX for recorded music product metadata - which is precisely the purpose of the DDEX ontology and messages.  Some people ("book" publishers) understandably want to use ONIX for music, and some ("record companies") will use DDEX for text,  and others (like Amazon.com or libraries using MARC want to integrate data from both sources). This kind of "mixed economy" will never change unless change itself stops because, as many have observed here, functional requirements vary endlessly at endless levels of granularity even where there is a reasonably common "world view".

 

On the issue of consistency of meaning: yes, everyone has private ontologies in their heads, and this affects both of the above approaches. It is good if we can agree standards, but there is no guarantee that my application of the standard is consistent with yours. To take an obvious example, the distinction between "Creator" and "Contributor" in Dublin Core hinges on the interpreter's judgement about what is a major or a minor contribution to the making of something , and there is no objective answer to that, so we will get a degree of semantic mismatch in the translation between different domains/systems/users. The risk of that can never be eliminated. Anyone looking for evidence of the existence of a plethora of private (and not always self-consistent) ontologies need look no further than our own discussions in this forum.

 

Where a mapping is involved, there is a second (and more dangerous) layer of possible mismatch between the terms of the ontologies/schemas themselves in the mapping. This is where the value of having explicit agreement between those responsible for the mapped schemas/ontologies is I think increasingly being recognized, but that also does no more than minimise the risk, as an explicit agreement on mapping is no guarantee of semantic identity (no doubt we are all familiar with variations on the saying "You may think you understand what I said, but what you don't understand is that I did not say what I meant").

 

Risk of semantic mismatch becomes greater the more steps a mapping goes through. In a "hub and spoke" approach such as we are taking with the VMF, a mapping between schemas A and B is dependent on intermediate mappings with a third schema rich enough to accommodate the semantics of both (in this case VMF), so for A and B to be confident, they need the assurance that these other mappings have been done to the same care and standard that their own has. Again this can never be guaranteed, but an explicit authorization process mitigates the risk.  In the VMF intiative we have not put any such process in place yet, but we have agreed it is essential going forward if anyone is going to commit to relying upon the mappings: I think such processes are likely to become more common if automated data interoperability between different schemas and ontologies is to be trusted outside of our widespread reliance on standard identifiers, date-times and numbers. 

 

I had meant to notify this forum of VMF some time ago - I note that a couple of other ontolog contributors are attending our event at the British Library in ten days and I look forward to meeting them there.  This first step of VMF, which maps a range of content industry/bibliographic vocabularies to a common core ontology, is funded by the UK's JISC grant body and the initial results in the form of an RDF/OWL ontology should be freely available before the end of the year.  see http://cdlr.strath.ac.uk/VMF/index.htm. It's a modest start to general semantic interoperability in this area, and it's not rocket science, but it has support from all the mapped schemes and we will welcome any and all constructive comments or suggestions after its initial publication.

 

Godfrey

 

Godfrey Rust
Chief Data Architect
Rightscom/Ontologyx
Linton House LG01
164/180 Union Street, London SE1 0LH
www.rightscom.com
Direct +20 8579 8655
Rightscom Office +20 7620 4433
Mobile 07967 963674

 

----- Original Message -----

From: Burkett, William [USA]

To: [ontolog-forum]

Sent: Friday, October 30, 2009 1:09 PM

Subject: Re: [ontolog-forum] Just What Is an Ontology, Anyway?

>MW: Ontology is about modeling  how things are in the world, not how  we talk about them.

 

But Matthew: You can?t get away from using natural language to create and talk about ontologies ? from which I conclude that ontologies naturally inherit the vaguarities of language.  The ideal that you?re striving for - *a* ontology for *the* real world just isn?t possible because (1) we?re bound by and limited by the natural languages we use to communicate and (2) software applications need data tailored to their particular purposes, which means localized semantics, which means applications developers don?t want to (or simply won?t) be hamstrung by a corporate ontology.

 

Besides: who is any of us to say how things *are* in the world.  All we can realistically do is express our view of them.

 

>MW: The alternative is to document all the individual ontologies and map between them. This is horrendously expensive,

>so a much cheaper alternative is to sit down and agree to use one in a particular way, which may not be the way that any

>of the individuals actually sees things, but at least is clear.

 

It?s not as expensive as you might think.  The mapping and translation simply hasn?t received the clear focus and attention it really needs and deserves; once it does, mapping and translation will be a lot easier and cheaper and adaptable.  (And if you use STEP as an example of what you think mapping is, can, or should be, I?d suggest thinking again.)  I personally think it?s more expensive in the long run to continually haggle over the meaning and evolution of a common model ? how much as STEP cost over these years?  (A horrendous amount, I dare say.)

 

>This is at least analogous to the situation you find when an Italian, and Frenchman,

>a German and a Spaniard have a business meeting. They speak English, rather than each have 3 translators.

 

It is analogous, yes, but there are three translators: each one has a personal translator in his head that converts the language to his internal meaning.

 

You can?t get away from the fact of ?personal ontologies?, or at least the personal, idiosyncratic views of the world embodied in databases.  Some kind of mapping and translation will always be necessary and I think robust adaptable solutions need to make it an essential element of the solution architecture.

 

Bill

 

 

 

From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Matthew West
Sent: Friday, October 30, 2009 2:35 AM
To: '[ontolog-forum] '
Subject: Re: [ontolog-forum] Just What Is an Ontology, Anyway?

 

Dear Rich,

 

Copied from below:

 

But then how do we account for the diverse viewpoints going into the system from multiple users?  We all agree that each user has a unique ontology of her personal world.  We know that subjectivity gets squeezed into the tightest databases with the strictest controls.  

 

MW: You  can?t. And in fact the problem really is just how do you impose sufficiently strict controls such that the range of meaning is sufficiently small that sufficiently accurate communication is possible.

 

You must.  As John Sowa is fond of saying, people play language games. 

 

MW: But computers don?t.

 

Those games are more complicated than we can decipher from signs alone.  So one enterprise level purpose of each subjective personal ontology is to ?correct? the personal viewpoint, projecting it back into the enterprise ontology.  

 

MW: This is essentially the process of agreeing the enterprise ontology, or aligning with it.

 

But note that if you project the disjunction of all personal ontologies to make up the enterprise ontology, you have to match common items shared among personal ontologies.

 

MW: I don?t know anyone who would do it like that. Much more likely is that a few people determine the enterprise ontology, and then others are left with aligning their own viewpoint with it.

 

For example, probably most or all normal English speakers think of fluids in one way, solids in another and gases in a third.  The English language reflects the way we talk about the things belonging to these different classes.

 

MW: One has to be very careful about this. Language includes lots of old ways of thinking about things that are not accurate. Ontology is about modeling how things are in the world, not how  we talk about them.

 

So there is clearly a linguistic common ontology of objects and classes that constitutes everyday usage.  

 

MW: No there is not, because with everyday language you can express any of the ontologies you might find. Words have such a variety of usages, that it can be difficult to accurately determine the meaning of words out o f context, and sometimes even in context.

 

That can be part of the enterprise ontology.  But its part of EVERY language competent ontology.  

 

MW: I?ve no idea what that might be. As I said above, ontology is about the things in the world, not how we talk about them.

 

So the enterprise ontology also includes things specific to the objects about which that enterprise is concerned.

 

Leading to the conclusion that the enterprise ontology will have to be multilayered, scalloped like a 50?s hot rod into component ontologies for each viewpoint and each group of viewpoints.  

 

MW: Well yes you can do that (maybe), but at prohibitive expense because of the interfaces between vewpoints, so I doubt if anyone will. This is back to why a meeting of a French, Italian, German and Spanish people will conduct business in English.

 

Regards

 

Matthew West                           

Information  Junction

Tel: +44 560 302 3685

Mobile: +44 750 3385279

matthew.west@xxxxxxxxxxxxxxxxxxxxxxxxx

http://www.informationjunction.co.uk/

http://www.matthew-west.org.uk/

 

This email originates from Information Junction Ltd. Registered in England and Wales No. 6632177.

Registered office: 2 Brookside, Meadow Way, Letchworth Garden City, Hertfordshire, SG6 3JE.

 

 

 

JMHO,

-Rich

 

Sincerely,

Rich Cooper

EnglishLogicKernel.com

Rich AT EnglishLogicKernel DOT com

---

From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Matthew West
Sent: Thursday, October 29, 2009 12:35 AM
To: '[ontolog-forum] '
Subject: Re: [ontolog-forum] Just What Is an Ontology, Anyway?

 

Dear Rich,

 

Looks like I missed something here.

 

 

So it looks like the consensus among those in this discussion is:

 

An ontology is a collection of

 

classes, each with possibly unique property values;

 

a few constant instances (e.g., equilateral triangle = special instance of generalized triangle, etc);

 

I?m not sure that I would see equilateral triangle as an instance. Surely there is more than one of them? On the other hand it could be a subtype of generalized triangle (some triangles are equilateral).

 

On the other hand there are plenty of individuals I might want in my ontology. If I want to define a class of ?Ford Motor Car? in my ontology  then it is useful to be able to have the individual ?Ford Motor Company? in my ontology so that I can make the restriction class.

 

and

logical relationships among the classes and instances.

 

And nothing else.  If that satisfies everyone, then any operational system would require more than just an ontology.  It would also require that information nobody seems to want to call ontological, like the specific employees in the employee table.  

 

If we accept this definition among the group of us, an ontology with a database to back it would be about the simplest semantic system I can imagine being useful.  The database would store the instance data beyond the ontology, but the ontology would define the classes, properties and relationships among the entities.  

 

MW: It will probably replicate much of the ontology too. It might be better to think of the ontology as an abstraction of the database, with some rules added, so that you can e.g. check the consistency of the database.

 

But then how do we account for the diverse viewpoints going into the system from multiple users?  We all agree that each user has a unique ontology of her personal world.  We know that subjectivity gets squeezed into the tightest databases with the strictest controls.  

 

MW: You  can?t. And in fact the problem really is just how do you impose sufficiently strict controls such that the range of meaning is sufficiently small that sufficiently accurate communication is possible.

 

So how do we account for personal ontologies in a semantic system?

 

MW: The alternative is to document all the individual ontologies and map between them. This is horrendously expensive, so a much cheaper alternative is to sit down and agree to use one in a particular way, which may not be the way that any of the individuals actually sees things, but at least is clear. This is at least analogous to the situation you find when an Italian, and Frenchman, a German and a Spaniard have a business meeting. They speak English, rather than each have 3 translators.

 

So I suggest we return to the real world.

 

Regards

 

Matthew West                           

Information  Junction

Tel: +44 560 302 3685

Mobile: +44 750 3385279

matthew.west@xxxxxxxxxxxxxxxxxxxxxxxxx

http://www.informationjunction.co.uk/

http://www.matthew-west.org.uk/

 

This email originates from Information Junction Ltd. Registered in England and Wales No. 6632177.

Registered office: 2 Brookside, Meadow Way, Letchworth Garden City, Hertfordshire, SG6 3JE.

 

______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email
______________________________________________________________________

---

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/ 
Config Subscr: 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 join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: 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 join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (01)