ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Is Philosophy Useful in Software Engineering Ontolog

To: "'[ontolog-forum] '" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Bruce Schuman" <bruceschuman@xxxxxxx>
Date: Mon, 22 Jun 2015 08:51:07 -0700
Message-id: <007a01d0ad03$40ddcfd0$c2996f70$@net>

Thanks.  This is funny.  I was just looking at this issue this morning from a very broadly "inspired" point of view.  I guess my short answer is -- maybe freshmen know something their jaded burned-out spirit-damaged tenure-desperate elders do not.  Break those beautiful wild horses and make them behave....

 

I used to have this general thesis -- that if you wanted to get into the real creative juice in computer science -- read the early work of the prime developers -- because their first books (usually based on their PhD thesis) contain their most seminal ideas. After they get "successful" and their spirit gets chewed up by departmental politics,  funding issues and industry pressures, and they decide what they really want in life is a comfortable and secure academic position, their ideas become boring (not to mention myopic and possibly gutless).  I used to think grad students were much more creative than their professors…

 

So, it depends on what you want to do.  For me, philosophy provides a broad perspective, within which particular ambitions get located. 

 

Here's a gorgeous slide-share presentation on "Logos" -- the history of the idea, with luscious graphics:

 

http://www.slideshare.net/Aesthetics_Art_Philosophy/a-history-of-an-idealogos-and-the-theory-of-the-original-image-uc-thought

 

The visionaries -- like anthropologist Gregory Bateson -- ask the question: "what is the pattern the connects?"  The philosophers with vision and courage explore that question.

 

***

 

Just a couple of days ago, I got into the very rich synoptic resources on Wikipedia to review related questions on "levels of language", because I want to see a smoothly integrated spectrum across these levels -- since I happen to believe it's possible, following stipulative definition (rather than empirical observation):

 

https://en.wikipedia.org/wiki/Programming_language - which does note in the section on taxonomies "There is no overarching classification scheme for programming languages" -- suggesting that at least for now, no clear criteria for approaching that subject is widely presumed

 

https://en.wikipedia.org/wiki/Low-level_programming_language

 

https://en.wikipedia.org/wiki/High-level_programming_language

 

And a page that opens up some keys to linking these things that I think could become quite powerful:

 

https://en.wikipedia.org/wiki/Abstraction_(computer_science)

 

This is not an eccentric or non-standard idea.  It's the basis of this entire industry, and this thought ought to be front and center:

 

"Computer science commonly presents levels (or, less commonly, layers) of abstraction, wherein each level represents a different model of the same information and processes, but uses a system of _expression_ involving a unique set of objects and compositions that apply only to a particular domain. Each relatively abstract, "higher" level builds on a relatively concrete, "lower" level, which tends to provide an increasingly "granular" representation. For example, gates build on electronic circuits, binary on gates, machine language on binary, programming language on machine language, applications and operating systems on programming languages. Each level is embodied, but not determined, by the level beneath it, making it a language of description that is somewhat self-contained."

 

I was looking at this page from the point of view of Ed Lowry's "microstructure" issues, particularly guided by the section on "control abstraction"- which goes into the exact linearly-sequential detail by which an _expression_ like  [ a := (1 + 2) * 5 ] gets evaluated.  This is foundational in universal terms.

 

It might be possible to make the case that the entire structure of human thinking – ranging from empiricism to holism – or from computer programming to philosophy -- can be spanned across a system of “levels” like this.

 

***

 

Bruce:

> so that universally-meaningful concepts like “identity” or “comparison”

> or “similarity” or “difference” or “analogy” can be given exact and

> very simple definitions with some hope of becoming consensual...

 

John:

That's not possible.  The reason why they're "universally meaningful"

is that they're vague, context-dependent, and sufficiently flexible that they can be adapted to any subject of any kind.

 

Fundamental principle:

 

  * If you're developing a normative ontology for a specific project

    or family of closely related projects, you can state precise

    definitions for all applications in that field.

 

  * But if you're trying to understand natural language on any

    topic, you can't be "Ontoclean" -- because NLs must be adaptable

    to anything.  Any ontology useful for NLP must be "Ontodirty".

 

Bruce:

I understand what you mean.  I don’t disagree with that.  To define these terms with greater universality requires a different kind of framework, and yes, it would be stipulative – where the meanings of terms are intentionally assigned, like the creation of a programming language – and not observation or empirical, like study of some particular specialized NLP domain – though there could be negotiated factors involved.   But – speaking cautiously -- I am not persuaded that there are no criteria for “better” or “best” ways of establishing consensual definitions across sectors.  I don’t accept the notion that groups in different sectors have no hope of understanding each, and I’d say “semantic ontology” could probably contribute constructive insights towards that end.

 

Bruce Schuman, Santa Barbara CA USA

http://networknation.net/vision.cfm

 

 

 

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

From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F Sowa

Sent: Monday, June 22, 2015 6:38 AM

To: ontolog-forum@xxxxxxxxxxxxxxxx

Subject: Re: [ontolog-forum] Is Philosophy Useful in Software Engineering Ontologies?

 

Leo, Bruce, and Tom,

 

I was tied up with some pressing deadlines.  But I'd like to make a few comments on some of the basic issues in this thread.

 

Leo

> Of course it helps if ontology engineers/implementers acquire a deep

> understanding of philosophy, and if philosophers acquire a deep

> understanding/practice of ontology engineers/implementers...

> 

> Understanding involves learning about both sides.

 

Yes, but a Phil 101 course for the engineers and a Software Engineering 101 course for the philosophers is far from sufficient.

Getting anything more for any but a tiny fraction of the people who need it is not going to happen.

 

Therefore, we have to face reality and work on Plan B.

 

Leo

> Once we have very solid NL technologies that can do semantic

> interpretation and mappings to ontologies in specific domains and

> contexts, then...

 

For unrestricted NLs, that will *never* happen.  I completely agree with Alan Perlis:  "You can't map informal language to formal language by any formal algorithm."

 

But with sufficient background knowledge and the ability to carry on a dialog to ask "Socratic" questions, it would be possible for a computer system to do quite a lot -- without going very far beyond the AI & NLP research available today.

 

Bruce

> so that universally-meaningful concepts like “identity” or “comparison”

> or “similarity” or “difference” or “analogy” can be given exact and

> very simple definitions with some hope of becoming consensual...

 

That's not possible.  The reason why they're "universally meaningful"

is that they're vague, context-dependent, and sufficiently flexible that they can be adapted to any subject of any kind.

 

Fundamental principle:

 

  * If you're developing a normative ontology for a specific project

    or family of closely related projects, you can state precise

    definitions for all applications in that field.

 

  * But if you're trying to understand natural language on any

    topic, you can't be "Ontoclean" -- because NLs must be adaptable

    to anything.  Any ontology useful for NLP must be "Ontodirty".

 

Tom

> I recommend an excellent article, "Categories", from the Stanford

> Encyclopedia of Philosophy.

> http://plato.stanford.edu/entries/categories/

 

I agree that the article is good of its kind.  But it's written by a philosopher for other philosophers, and it would not give Rich or any other software developer any useful guidelines.

 

In the following points, I'm not criticizing the author of that article -- I'm criticizing the entire academic tradition:

 

  1. The author starts with Aristotle, but does not mention that

     A's book on categories was part of the Organon (Instrument)

     for doing science.  A's father was a physician, from whom he

     learned the importance of careful observation and recording

     of details.

 

  2. Aristotle wrote more about biology than he did about metaphysics,

     and he did very careful experiments. From the very beginning, he

     developed his categories as an *applied* science.  Galen was very

     impressed by Aristotle, and he couched his vast writings on

     medicine in an Aristotelian framework.

 

  3. Section 1.6 has two paragraphs on "Categories in other disciplines"

     and just 3 references.  The first paragraph mentions "information

     systems" and cites another philosopher.  The second cites two

     cognitive scientists, but says nothing about linguistics and

     lexicography.

 

  4. Those are subjects that stimulated the work on categories by

     Leibniz and others in the 17th century, which had a strong

     influence on Kant in the 18th c.  That work in the 17th & 18th

     centuries had a strong influence on Roget in the 19th.  But

     philosophers have totally ignored lexicography (except for

     Peirce, who was an associate editor of the _New Century

     Dictionary_ -- that's one reason why his work is still at the

     cutting edge of research in AI & NLP).

 

  5. In Section 1.4 on "Contemporary category systems", the author

     mentions a few that happened to be developed in the late 20th c.

     Ingvar Johansson's system is the only one that would be useful

     as a basis for applied ontology.  Ingvar, by the way, did

     participate in the discussions in Ontolog Forum for a few months.

     There were some good discussions, but like Pat Hayes and others,

     he dropped out (and I don't blame him).  One point that's very

     important about IJ's system is that he explicitly recognizes

     intentionality as a fundamental category.

 

  6. But the worst feature of that article (and of a huge amount

     of the philosophical literature) is a total disregard of

     two fundamental issues:  (1) modern science, and (2) the

     sign systems by which all living things communicate.

     I admit that many philosophers do address those issues,

     but more often than not, they just try to develop an

     ontology of "common sense" -- which is usually a fossilized

     version of some obsolete version of science.

 

  7. Science and communication were fundamental for Aristotle.

     In addition to biology (which was his primary science),

     he addressed and named a large number of the "-ologies"

     people have been studying for over two millennia.

     And the first paragraph of A's book _On Interpretation_

     introduces the meaning triangle (often called Ogden and

     Richard's meaning triangle).  For more about that, see

     my article, "The role of logic and ontology in language

     and reasoning", http://www.jfsowa.com/pubs/rolelog.pdf

 

As I said above, I blame the entire academic tradition for these faults -- and it permeates *every* field, not just philosophy.

 

I have known several very good academics who did excellent work in relating their research to an interdisciplinary collaboration with other fields.  But they made the mistake of doing that

*before* they got tenure.

 

When explaining why they were fired, the head of the department would say things like "you didn't seem to be committed to the field"

or "your work was outside the mainstream" or "maybe you'd be happier doing something else" or, worst of all, "he inspires freshmen".

 

And by the way, enrollments in philosophy courses and departments has declined drastically during the past century.  You can see the effects in bookstores.  Just count the number of shelves devoted to "philosophy" in comparison to "new age" and other such topics.

 

Those other topics "inspire freshmen".  Any subject that doesn't inspire freshmen won't get any sophomores to enroll in courses.

 

John

_________________________________________________________________

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


_________________________________________________________________
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    (01)

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