Doug, Aldo, Kingsley, and Ed, (01)
DF
> both [the Cyc and OWL] expressions say that EVERY squid has exactly 10
> tentacles; they don't say that any squid exists. (02)
Sorry. I hastily wrote "there is" in my English sentence. We all make
such slips, as Aldo also did: "Please read 'max 4' not 'min 4' " (03)
Following are the correct translations for Simon's example: (04)
English: Every squid has as part exactly 10 tentacles.
CG: [Squid: @every]->(HasPart)->[Tentacle: {*}@10]. (05)
AG
> In OWL (serialised in Turtle):
>
> :Mary a [:hasChild max 4] ; :hasChild :Sue , :Bill , _:x .
> :Sue a :Girl . :Bill :hasTwin _:x .
>
> More readable than CG in my view ... (06)
Very nice. This is better than I had expected. For comparison, I
rearranged your 4 lines to 2 lines (and replaced 'min' with 'max'). (07)
English: Mary has less than 5 children, including a girl named Sue,
and a pair of twins, one of whom is named Bill." (08)
CG: [Person: Mary]->(HasChild)->[Person: {Sue, Bill, *x, *}@<5]
[Girl: Sue] [Person: Bill]->(HasTwin)->[Person: ?x] (09)
AG
> The problem is not DL OWL or others, but being iconic enough to support
> intuition. Usually NL works fine but in this case OWL/Turtle is more intuitive (010)
I agree that the English sentence is less systematic than either the CG
or the OWL/Turtle translation. But the readability of any formalism
depends heavily on prior experience in using it. (011)
KI
> Yes, but how does one lift your example into a HTTP network medium
> where documents create a hypermedia mesh of Linked Documents comprised
> of content that delivers Linked Data? (012)
The primary design goal for CGs is to serve as an intermediate language
between NLs and formal notations. A major use for VivoMind software is
to analyze documents with *no links* and *discover* the relevant links. (013)
For examples, see http://www.jfsowa.com/talks/goal7.pdf . None of the
documents analyzed for those applications had any annotations. We're
happy to use linked data, but our software can also create the links. (014)
EJB
> First, note that the above example uses two Classes: Person and Girl,
> and two objectProperties: HasChild and HasTwin. (015)
In CG notation, every concept box (or square brackets in the linear
format) has a *type field* and a *referent field* separated by ':'. (016)
The notation [Girl: Sue] declares 'Girl' to be the name of a monadic
relation that specifies the type, and 'Sue' to be the name of an
instance of type Girl. (017)
EJB
> In OWL, one would declare:
> - Girl to be subclassOf Person (which John omits) (018)
The CG was a direct translation of the English sentence. In mapping
any NL to any formalism, it's useful to look up words in a lexicon
that relates words to an ontology. But it's also possible to derive
a significant amount of the ontology and related information from
the NL documents themselves. (See the examples in goal7.pdf.) (019)
EJB
> perhaps the domain and range of HasChild and HasTwin to be restricted
> to Persons, and hasTwin to have maximum cardinality 1 and be symmetric
> (none of which John mentions). (020)
Yes. I assumed that subscribers to Ontolog Forum would know how to
define such things in whatever notation they prefer. (021)
EJB
> And OBTW, writing the cardinality constraint for Mary's children is
> a lot easier in OWL than in a FOL language like KIF. (022)
No! As I said in my previous note, OWL has a large amount of built-in
ontology that a "bare bones" logic such as EGs or CL does not. But you
can take a bare bones notation such as EG, add an appropriate ontology,
and define whatever "syntactic sugar" may be useful. Note that EGIF
is a subset of CGIF. See http://www.jfsowa.com/talks/egintro.pdf (023)
EJB
> For the purpose of reasoning, it is useful to sort out the information
> into the simplest axioms... (024)
There has been over two millennia of very sophisticated "sorting out"
of the many logical, ontological, and linguistic issues. For modern
symbolic logic, the emphasis on the foundations of math has ignored
many issues about more general ontologies. (025)
Doug Lenat gives every potential new hire a list of English sentences
to be translated to FOL. He apologized to one candidate, who taught
logic at a university. But he said that he asked everybody to take
the test. Unfortunately, the logic teacher failed the test. He said
"That's not what we teach our students." That is certainly true. (026)
John (027)
_________________________________________________________________
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 (028)
|