Dear Doug, Gary, Chris, (01)
Sorry for the delay in responding; I finally got
time to study the Cyc statements versus the
questions in the discussion below. Looking at the
CycL statements, (with intensity), it appears to
have a nicely structured type system, and seems to
solve all the problems mentioned in the
discussion. (02)
How difficult would it be to build a program that
would learn these kinds of distinctions from
available texts corpora? I.e., that organizations
are slightly different than individuals; that
certain kinds of organizations through their
representatives act similarly to other kinds, and
to otherwise develop structures that compact the
representation to a most dense order? (03)
It seems that the way in which words (such as
"organization", "organizations", "Bishopric" ...)
are used in context should provide some
statistical narrowing which could focus a learning
mechanism (like AIXI) to compact the
representation automatically. (04)
Of course, that approach has been tried in many
variants - unsupervised learning of vocabularies
from corpora - with varied results. But is there
an analysis of the structure of an arbitrary
corpus, using an arbitrary software tool that is
believed reliable by its history of past
production success? (05)
I am reading Graeme Hirst's excellent book
"Semantic Interpretation and the Resolution of
Ambiguity" which was recommended here a couple
weeks ago, and it points out some of the
structural ideas behind disambiguation very
clearly. Some of these ideas, along with software
tools to analyze n-grams and collocations of
various kinds, ought to at least leverage labor in
the analysis of corpora. Have there been any
landmark papers or studies or other descriptions
of such analyses? (06)
Thanks for any references,
-Rich (07)
Sincerely,
Rich Cooper
EnglishLogicKernel.com
Rich AT EnglishLogicKernel DOT com
9 4 9 \ 5 2 5 - 5 7 1 2 (08)
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On
Behalf Of doug foxvog
Sent: Wednesday, February 01, 2012 10:47 PM
To: 'Ontology Summit 2012 discussion';
[ontolog-forum]
Subject: Re: [ontolog-forum] [ontology-summit]
System Components (09)
> And if something (in Cyc?) is intangible, does
it have spatio-temporal
> location? (010)
If it is a #$SpatialThing-Localized, it has a
spatial location and a
temporal location, but a Platonic circle is a
spatial thing without a
location, while an agreement has a temporal
location but not a
spatial one. (011)
#$Situations, including #$Events are considered to
be intangible,
even if they have physical participants and
results. (012)
Does an organization, e.g., IBM have a spatial
location and extent?
Does it have a physical size, a physical shape, a
weight or mass?
It can take an action (through a representative)
any place that
representatives is. An organization certainly
has a temporal location.
I would suggest that it does not have a physical
extent or location.
Its representatives, employees, head office,
outlets, manufacturing,
and service facilities all have physical extents
and locations. The
location of my property or my representatives do
not define my
location -- why should it do so for an
organization? (013)
The same analysis holds for the Anglican Church or
any of its
suborganizations. I would hold that a bishopric
(as an organization
sole) is a suborganization of the church. When
the holder of the
office moves around, that shouldn't (imo) mean
that the office
moves around. (014)
> E.g. The Bishop of Liverpool was in Manchester
yesterday. (015)
This is where ontologies win over natural
language. I previously
said that we were discussing three things, an
organization sole
(the office of bishop, or bishopric), the person
holding that office,
and the situation of that person holding that
office (which can be
represented by a relation that holds during some
time interval
between the person and the office). The word
"bishop" can refer
to any of these inter-related things. (016)
In CycL:
First define your terms:
(isa bishopOf BinaryPredicate)
(arg1Isa bishopOf HumanAdult)
(arg1Isa bishopOf Bishop-Clerical)
(arg2Isa bishopOf Bishopric)
(genlInverse bishopOf hasLeader) (017)
(isa Bishopric ExistingObjectType)
(genls Bishopric ReligiousOrganization)
(genls Bishopric OrganizationSole) (018)
(isa AnglicanBishopric ExistingObjectType)
(genls AnglicanBishopric Bishopric) (019)
(isa Bishop-TheWord EnglishWord)
(singular Bishop-TheWord "bishop")
(denotation Bishop-TheWord Noun 0 Bishopric)
(denotation Bishop-TheWord Noun 1
Bishop-Clerical)
(denotation Bishop-TheWord Noun 2
Bishop-ChessPiece)
(denotation Bishop-TheWord Noun 3 bishopOf)
(compoundString Bishop-TheWord "of" Noun
bishopOf)
<other more complex NL assertions> (020)
Then define your data (in a different
microtheory):
(holdsDuring
(TimeIntervalBetweenFn (YearFn 1880)
Now-Indexical)
(isa BishopricOfLiverpool AnglicanBishopric)) (021)
(compoundString Bishop-TheWord "of Liverpool"
Noun
BishopricOfLiverpool ) (022)
(isa JamesStuartJones
(holdsDuring
(TimeIntervalBetweenFn (YearFn 1998)
Now-Indexical)
(bishopOf JamesStuartJones
BishopricOfLiverpool)) (023)
> Can something intangible do things? If so, is
there some physics to
> support this? (024)
An organization does things through its authorized
representatives.
An account automatically charges or grants
interest even though it
is not physical. (025)
> E.g. The Bishop of Liverpool said Mass in
Manchester yesterday. (026)
This English phrase means that the person who held
the office of
Bishop of Liverpool yesterday performed a specific
role in a specific
type of church service yesterday. After finding a
match for
(holdsDuring Yesterday-Indexical
(bishopOf ?SUBJECT BishopricOfLiverpool))
create an instance of #$ChurchService (or
#$Mass-Ritual if that
were ontologized), assert for that instance (here
called ?MASS)
(temporallySubsumes Yesterday-Indexical ?MASS)
(eventLocation ?MASS (TerritoryFn
CityOfManchesterEngland))
(providerOfService ?SUBJECT ?MASS) (027)
FWIW, I would consider an organization to be
intangible (see my
discussion of organizations above), but Cyc has
not defined it that way.
Cyc is not perfect -- nor am i. (028)
[Oops. I left out ontologizing the situation type
of someone holding
the office. But it's late. I'm going to post
this anyway. I hope you
get the idea anyway.] (029)
-- doug (030)
>
>> -----Original Message-----
>> From: ontology-summit-bounces@xxxxxxxxxxxxxxxx
[mailto:ontology-summit-
>> bounces@xxxxxxxxxxxxxxxx] On Behalf Of doug
foxvog
>> Sent: 01 February 2012 07:00
>> To: Ontology Summit 2012 discussion
>> Subject: Re: [ontology-summit] System
Components
>>
>> >>what happens when the next person gets
instantiated as the (same)
> bishop?
>>
>> > Here is one take on it (without mentioning
4D?) The role of bishop of
>> > Liverpool was used by Person A for
time-period 100.
>> > That role may be used by Person B for the
next time period 101. You
>> > may identify this triple relationship as you
wish.
>>
>> We are talking about three things, not two, the
corporation sole, the
>> role
> of
>> bishop, and the person who occupies the role of
bishop for a given
> bishopric.
>>
>> The Bishopric is an intangible individual (like
a corporation) that can
> own
>> property and has been granted authority to
perform certain kinds of
>> acts.
>>
>> A Person while performing the role of Bishop,
has the derived authority
>> to
>> perform the actions which the Bishopric is
permitted. This is the same
>> as
> any
>> other corporate officer being permitted to act
in the name of her
> corporation
>> when authorized.
>>
>> -- doug f
>>
>>
>> > Regards,
>> > Gary
>> >
>> >
>> > On Tue, Jan 31, 2012 at 5:57 PM, Chris
Partridge
>> > <partridgec@xxxxxxxxxxxxxxx
>> >> wrote:
>> >
>> >> It looks like you are describing something
similar to a qua object,
>> >> which has no identity.****
>> >>
>> >> Not sure we are talking about the same
thing. Bishop in this case is
>> >> a particular bishop, such as the 'Bishop of
Liverpool' - a
>> >> corporation sole, a legal person. If we take
the approach you
>> >> suggest, what happens when the next person
gets instantiated as the
>> >> (same) bishop. In what sense is it the
'same' bishop - the bishop
>> >> continues to own the land, etc.****
>> >>
>> >> Wouldn't Occam's razor prefer a simple
agent.****
>> >>
>> >> ** **
>> >>
>> >> Regards,****
>> >>
>> >> Chris ****
>> >>
>> >> ** **
>> >>
>> >> *From:*
ontology-summit-bounces@xxxxxxxxxxxxxxxx [mailto:
>> >> ontology-summit-bounces@xxxxxxxxxxxxxxxx]
*On Behalf Of *Gary
>> >> Berg-Cross
>> >> *Sent:* 31 January 2012 21:47
>> >>
>> >> *To:* Ontology Summit 2012 discussion
>> >> *Subject:* Re: [ontology-summit] System
Components****
>> >>
>> >> ** **
>> >>
>> >> Chris,
>> >>
>> >> Chris noted>I find it odd that a qua object
can own property, have
>> >> responsibilities, declare war (Monarchs),
etc. - which the person
>> >> cannot do.
>> >>
>> >> Thematic roles may be thought of as a set of
types when objects
>> >> participate in processes. There is a
general notion of the role, but
>> >> it gets instantiated in an object, such as a
person who has a bishop
>> >> role and thus participates in the bishop
processes.****
>> >>
>> >> Gary Berg-Cross, Ph.D. ****
>> >>
>> >> gbergcross@xxxxxxxxx ****
>> >>
>> >>
http://ontolog.cim3.net/cgi-bin/wiki.pl?GaryBergCr
oss****
>> >>
>> >> NSF INTEROP Project ****
>> >>
>> >>
http://www.nsf.gov/awardsearch/showAward.do?AwardN
umber=0955816
>> ****
>> >>
>> >> SOCoP Executive Secretary****
>> >>
>> >> Knowledge Strategies ****
>> >>
>> >> Potomac, MD****
>> >>
>> >> 240-426-0770****
>> >>
>> >> ** **
>> >>
>> >> On Tue, Jan 31, 2012 at 1:05 PM, Chris
Partridge <
>> >> partridgec@xxxxxxxxxxxxxxx> wrote:****
>> >>
>> >> You will find references to 'bishop' in the
article.****
>> >>
>> >> I understand that Nicola and Giancarlo (and
others) would regard
>> >> 'bishop'
>> >> as a role, qua object, etc.****
>> >>
>> >> Historically this 'role' was introduced to
ensure that legal title to
>> >> property given to the church remained with
the bishop (corporation
>> >> sole) rather than inherited by his kin. It
was then extended to
>> >> monarchs.****
>> >>
>> >> I find it odd that a qua object can own
property, have
>> >> responsibilities, declare war (Monarchs),
etc. - which the person
>> >> cannot do.****
>> >>
>> >> I also cannot see how this would work if the
corporation sole did not
>> >> have some kind of identity.****
>> >>
>> >> ****
>> >>
>> >> *From:*
ontology-summit-bounces@xxxxxxxxxxxxxxxx [mailto:
>> >> ontology-summit-bounces@xxxxxxxxxxxxxxxx]
*On Behalf Of *Jack Ring
>> >> *Sent:* 31 January 2012 13:24****
>> >>
>> >>
>> >> *To:* Ontology Summit 2012 discussion
>> >> *Subject:* Re: [ontology-summit] System
Components****
>> >>
>> >> ****
>> >>
>> >> I do not find 'role' mentioned in the given
URL regarding 'sole'****
>> >>
>> >> Elsewhere in widipedia is "role posits the
following about social
>> >> behaviour: #3. Roles are occupied by
individuals, who are called
>> >> actors<http://en.wikipedia.org/wiki/Actors>
>> >> ."****
>> >>
>> >> Overall It seems that role specifies a
location in a coordinate
>> >> system, not an operator such as actor.****
>> >>
>> >> ****
>> >>
>> >> Seems to me that the ****
>> >>
>> >> On Jan 31, 2012, at 1:50 AM, Chris Partridge
wrote:****
>> >>
>> >> ** **
>> >>
>> >> One sense, for another see e.g.
>> >>
http://en.wikipedia.org/wiki/Corporation_sole****
>> >>
>> >> 'role' covers a wide variety of meanings****
>> >>
>> >> ****
>> >>
>> >> *From:*
ontology-summit-bounces@xxxxxxxxxxxxxxxx [mailto:
>> >> ontology-summit-bounces@xxxxxxxxxxxxxxxx]
*On Behalf Of *Jack Ring
>> >> *Sent:* 30 January 2012 23:33
>> >> *To:* Ontology Summit 2012 discussion
>> >> *Subject:* Re: [ontology-summit] System
Components****
>> >>
>> >> ****
>> >>
>> >> According to David Taylor, Object Technology
for Managers, role is
>> >> simply an authorization to act, not the
operator that acts. Operators
>> >> inherit roles then process operands
accordingly. ****
>> >>
>> >> ****
>> >>
>> >> On Jan 30, 2012, at 3:50 PM, Nicola Guarino
wrote:****
>> >>
>> >>
>> >>
>> >> ****
>> >>
>> >> Dear Matthew, ****
>> >>
>> >> ****
>> >>
>> >> just a few clarifications
concerning my lab's work. Note
>> >> that I am just trying to catch up with the
(main points of the)
>> >> discussion, and I am probably missing many
things. I look forward to
>> >> seeing the discussion synthesised somewhere,
in order to allow
>> >> everybody to understand how we
>> >> progress.****
>> >>
>> >> ****
>> >>
>> >> Matthew West writes (answering to Giancarlo
Guizzardi):****
>> >>
>> >> An alternative to this issue can be thought
of by considering qua
>> >> individuals
>> >>
(e.g.http://www.loa.istc.cnr.it/Papers/KR04MasoloC
.pdf)****
>> >>
>> >> ****
>> >>
>> >> MW: This is very similar to the 4D, but is
relatively opaque, and
>> >> gives more individuals than if you adopt
extensional identity in 4D.
>> >> In this case playing multiple roles
simultaneously does not give
>> >> multiple states, but one state playing
multiple roles. A bit more
>> >> elegant I think.****
>> >>
>> >> ****
>> >>
>> >> or perspectiles
>> >>
http://www.loa.istc.cnr.it/Papers/BottazziFerrario
PerspectilesEuroCog
>> >> Sciv.pdf
>> >> ). ****
>> >>
>> >> ****
>> >>
>> >> MW: This seems to generalise the idea above
a bit. One problem I have
>> >> with both of these is that (if I understand
it correctly) they treat
>> >> social and other roles as purely classes.
This gives me a problem if
>> >> I want to shake the hand of the president,
or start P101, because
>> >> classes are abstract, and these are just
things you can't do to them.
>> >> This is central to what I find
unsatisfactory with these kinds of
>> >> approaches. The situation is confused by
there being several
>> >> different meanings to role, from the
participant role in an activity
>> >> or state, to the component in a system, or
social role with
>> >> significant differences in character between
them.****
>> >>
>> >> ****
>> >>
>> >> The second paper is still work in progress,
while the first one is
>> >> more established. In both cases, however,
for sure the approach does
>> >> not only admit roles as "pure classes", and
new kinds of individuals
>> >> are introduced.
>> >> I defend a similar, although slightly
different approach in the paper
>> >> below, which explicitly considers the parts
replacement problem
>> >> (among other things) by introducing the
notion of a "virtual
>> >> individual" (NOTE
>> >> -
>> >> this is still a draft - comments
welcome):****
>> >>
>> >> ****
>> >>
>> >>
http://dl.dropbox.com/u/7903842/Guarino-Functional
PartsFunctionalRole
>> >> s.pdf
>> >> ****
>> >>
>> >> ****
>> >>
>> >> Best,****
>> >>
>> >> ****
>> >>
>> >> Nicola****
>> >>
>> >>
>> >>
>>
__________________________________________________
_______________
>> >> Msg Archives:
http://ontolog.cim3.net/forum/ontology-summit/
>> >> Subscribe/Config:
>> >>
http://ontolog.cim3.net/mailman/listinfo/ontology-
summit/
>> >> Unsubscribe:
>> >>
mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx<onto
logy-summit-
>> leave@o
>> >> ntolog.cim3.net> Community Files:
>> >>
http://ontolog.cim3.net/file/work/OntologySummit20
12/
>> >> Community Wiki:
>> >>
http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySu
mmit2012
>> >>
>> >> Community Portal:
http://ontolog.cim3.net/wiki/****
>> >>
>> >> ****
>> >>
>> >>
>> >>
>>
__________________________________________________
_______________
>> >> Msg Archives:
http://ontolog.cim3.net/forum/ontology-summit/
>> >> Subscribe/Config:
>> >>
http://ontolog.cim3.net/mailman/listinfo/ontology-
summit/
>> >> Unsubscribe:
>> >>
mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx<onto
logy-summit-
>> leave@o
>> >> ntolog.cim3.net> Community Files:
>> >>
http://ontolog.cim3.net/file/work/OntologySummit20
12/
>> >> Community Wiki:
>> >>
http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySu
mmit2012
>> >>
>> >> Community Portal:
http://ontolog.cim3.net/wiki/ ****
>> >>
>> >> ****
>> >>
>> >>
>> >>
>> >>
>>
__________________________________________________
_______________
>> >> Msg Archives:
http://ontolog.cim3.net/forum/ontology-summit/
>> >> Subscribe/Config:
>> >>
http://ontolog.cim3.net/mailman/listinfo/ontology-
summit/
>> >> Unsubscribe:
mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
>> >> Community Files:
>> >>
http://ontolog.cim3.net/file/work/OntologySummit20
12/
>> >> Community Wiki:
>> >>
http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySu
mmit2012
>> >> Community Portal:
http://ontolog.cim3.net/wiki/****
>> >>
>> >>
>> >>
>> >>
>> >> -- ****
>> >>
>> >>
>> >>
>>
__________________________________________________
_______________
>> >> Msg Archives:
http://ontolog.cim3.net/forum/ontology-summit/
>> >> Subscribe/Config:
>> >>
http://ontolog.cim3.net/mailman/listinfo/ontology-
summit/
>> >> Unsubscribe:
mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
>> >> Community Files:
>> >>
http://ontolog.cim3.net/file/work/OntologySummit20
12/
>> >> Community Wiki:
>> >>
http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySu
mmit2012
>> >> Community Portal:
http://ontolog.cim3.net/wiki/
>> >>
>> >>
>> >
>> >
>> > --
>> > Gary Berg-Cross, Ph.D.
>> > gbergcross@xxxxxxxxx
>> >
http://ontolog.cim3.net/cgi-bin/wiki.pl?GaryBergCr
oss
>> > NSF INTEROP Project
>> >
http://www.nsf.gov/awardsearch/showAward.do?AwardN
umber=0955816
>> > SOCoP Executive Secretary
>> > Knowledge Strategies
>> > Potomac, MD
>> > 240-426-0770
>> >
>> >
>>
__________________________________________________
_______________
>> > Msg Archives:
http://ontolog.cim3.net/forum/ontology-summit/
>> > Subscribe/Config:
>> >
http://ontolog.cim3.net/mailman/listinfo/ontology-
summit/
>> > Unsubscribe:
mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
>> > Community Files:
http://ontolog.cim3.net/file/work/OntologySummit20
12/
>> > Community Wiki:
>> >
http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySu
mmit2012
>> > Community Portal:
http://ontolog.cim3.net/wiki/
>> >
>>
>>
>>
>>
__________________________________________________
_______________
>> Msg Archives:
http://ontolog.cim3.net/forum/ontology-summit/
>> Subscribe/Config:
>
http://ontolog.cim3.net/mailman/listinfo/ontology-
summit/
>> Unsubscribe:
mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
>> Community Files:
http://ontolog.cim3.net/file/work/OntologySummit20
12/
>> Community Wiki:
>>
http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySu
mmit2012
>> Community Portal: http://ontolog.cim3.net/wiki/
>
> (031)
__________________________________________________
_______________
Message Archives:
http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr:
http://ontolog.cim3.net/mailman/listinfo/ontolog-f
orum/
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?WikiHomePa
ge#nid1J (032)
_________________________________________________________________
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 (033)
|