|
I agree with you Steve, and think I'm guilty of wording things in a way that makes that suggestion. I do think it is true, though, that different languages tend to increase or decrease the frequency of different kinds of errors. The languages themselves can do this; introduction and training approaches and materials can do this, and the interactions between a given language and human cognitive tendencies and/or more variable, customary practices can do this. Particular errors can increase or decrease because of modelling approaches suggested, or because of modelling approaches made easy or harder, or more or less obvious.
I meant to say that, while I don't know what combination of these causal routes is behind the correlation, my observation has been that the noun-centric modeling error happens more often in contexts in which OWL is the dominant or only ontology language than in projects using other languages and/or there is no experienced hand leading methodology on the ground. And the error seems, anecdotally, to have increased in frequency as these two types of situations increased in frequency.
Only correlation and anecdote, and only meant to counter the suggestion (unmeant, as it turned out), that the noun-centric error was rooted in CS traditions.
Best,
Amanda
On Sun, Apr 29, 2012 at 1:44 PM, Steve Ray <steve.ray@xxxxxxxxxxxx> wrote:
It seems to me that this conversation is implying that because people are creating poor models using OWL, that OWL is therefore a poor modeling language. All of the problems raised in the wine example quoted can be properly modeled in OWL. So while I agree that OWL is less expressive than FOL, I think we should be careful to distinguish issues with a language from issues with people who use the language.
I’m no expert in FOL modeling, but I’m confident there are plenty of examples of horrible modeling in it as well.
- Steve
Steven R. Ray, Ph.D.
Distinguished Research Fellow Carnegie Mellon University
NASA Research Park Building 23 (MS 23-11)
P.O. Box 1
Moffett Field, CA 94305-0001 Email: steve.ray@xxxxxxxxxx
Phone: (650) 587-3780
Cell: (202) 316-6481
Skype: steverayconsulting 
As one who fights the battle against the axiomatization of nonsense in the financial services domain on a daily basis, I can assure you that the OWL partisans are only picking up where their UML brethren left off.
On Sat, Apr 28, 2012 at 5:16 PM, Amanda Vizedom <amanda.vizedom@xxxxxxxxx> wrote:
I agree with you that the noun-centric approach is an error. I haven't experienced it as a general computer science approach though. I really didn't see it in action when I working as an ontologist in organizations or projects that were focused on, or grew out of, AI, NLP, simulation, or broad-scope IR. I really began to see it more recently as (1) OWL became more popular and (2) More projects started up that use ontology but had no very experienced ontologists involved at the technical, ground level. I think that OWL and/or the most popular training materials for it, tend to exacerbate this kind of error. In these project where methodology is established without an experienced hand to guide, people will often grab onto something like parts of speech that is familiar, but not actually a good proxy for ontological significance.
Amanda
On Sat, Apr 28, 2012 at 16:31, William Frank <williamf.frank@xxxxxxxxx> wrote:
I am especially interested in your mentioning Donald Davidson as a source for "Treating Events as First Class Entities."
Computer science mostly seems to have taken the position that only the entities that are primarily nouns in Indo-European lanaguages should be regared as entitites.
On grounds of practicality, this is unlikely a good approach, for, as what started this thread, the organization of the things that we ask services to do (peform actions) is almost as important as what we ask them to do it to (some passive data entitites)
On grounds of fundamentals, which when ignored, tend to lead to practical problems later, it is unlikely that just because some group of people tend to *view* certain sequences of observable phenomina as representing a "Thing" and others of them as representing an "event" is not, to me, a difference in kind. In both cases, we can make more than one observation over time, and say, that is the same walk that Sebasitan has been on for the last hour, this is the same hurricane he was walking in, as easily as that is the hat he is wearing. I do seem to vaguely recall that there are northwest North American Indian languages that do not make any distinction between these kinds of cases. "Being John Malkovich" and "being the eating of a salmon by a bear" are, in this language, each regarded as "things" or alternatively, are both regarded as "events" (since no difference is made between the two). Both events take some time, one event simply take much longer than the other. Perhaps the difference is really only in the number of relations that are required to define the event as it is thought of. Three for the eating, only one for John.
On Sat, Apr 28, 2012 at 1:03 PM, Obrst, Leo J. <lobrst@xxxxxxxxx> wrote:
Heh - Haven't heard that mentioned in a while;. I always had the impression that KQML sort-of fizzled out; I don't think anyone was working with it when I was at EIT 94-95.
I am notifying the subscribers of this list of a few sources that might be useful to them (there are several different topics that have come up on this thread by typing <<
The foundational work on speech acts is John Austin's "How to do things with words." (Austin 1962). John Searle aso wrote on the subject.
An important work on actions and events is Davidson's "Essays on Action and Events". (Davidson 1980). Good commentaries can be found in LePore, E. and McLaughlin, B. P. , eds. (1988).
The Davidsonian approach treats events as first class entities. We can represent the meaning of "Sebastian walked in Bologna at midnight" as a series of statements:
There is a Walking, Walk1. A walker in Walk1 is Sebastian. A location of Walk1 is Bologna. A time of Walk1 is midnight.
An advantage of using this approach as compared to using predicates with an argument for each adverbial modifier of the walk is that the number of such modifiers, (and hence the number of predicates of different arity) can be unbounded, and it is not clear how to generate the necessary entailments - for example, that someone walked in Bologna at midnight; that Sebastian walked in Bologna, etc.
Wordnet - http://wordnet.princeton.edu/ - contains a hierarchy of verbs; however Wordnet is not not an ontology, and there are known issues in the hierarchy (see e.g. Richens 2008).
Cyc uses a Davidsonian model of events; Austin, J. L. (1962). How to do things with words : the William James lectures delivered at Harvard University in 1955. Clarendon Press, Oxford.
Davidson, D. (1980). Essays on actions and events. Clarendon Press ;, Oxford. LePore, E. and McLaughlin, B. P. eds. (1988). Action and events : perspectives on the philosophy of Donald Davidson. B. Blackwell,, Oxford, UK ;New York, NY, USA.
Richens, T. (2008). Anomalies in the wordnet verb hierarchy. In Proceedings of the 22nd International Conference on Computational Linguistics-Volume 1, pages 729–736. Association for Computational Linguistics. Available at: http://aclweb.org/anthology-new/C/C08/C08-1092.pdf
>>
--
William Frank
413/376-8167
This email is confidential and proprietary, intended for its addressees only.
It may not be distributed to non-addressees, nor its contents divulged,
without the permission of the sender.
_________________________________________________________________
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
--
Kevin P. Tyson
Kevin.Tyson@xxxxxxxxx
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2012.0.1913 / Virus Database: 2411/4966 - Release Date: 04/29/12
_________________________________________________________________
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)
|