RE: SUO: Organization
Pat,
Lots of good sense.
My comments:
PC>The roles and positions would be related to the
Collection "OrganizationMembers" by another predicate, e.g.
as in TOVE:
(filledBy OrganizationPosition OrganizationMember)
CP> If positions are to be treated as kinds of organization (units) /
persons we need a filled by relation. I'm not sure it is organization
members that should be the 'role' in this relation. Some 'roles' such as
employment create a membership without involving a position - others like
appointment of non-executive directors bundle up creating the membership and
filling the position. So maybe we should substitute Person for Organisation
Member. Noting that person covers both humans and organizations.
PC>For example, it appears that
in TOVE an outside contractor can fill an Organization-Role.
TOVE has an Organization-Entity concept that has contractors
as subclass. I would prefer to keep contractors separate
from members, so we might need a more general term such as
the TOVE Organization-Entity which would subsume OrganizationMember
CP>I agree this is odd. However, I think the legal notion of outside
contractor is often driven by tax avoidance rather than the content of the
work. Most of us know people who have been (legally) contractors at
organizations for longer than many permanent employees - and feel and are
recognized as 'part' of the organization.
It seems to me that you are right to say that legally outside contractors
are not members, but that socially they are.
No more comments.
Regards
Chris
-----Original Message-----
From: pcassidy@bellatlantic.net [mailto:pcassidy@bellatlantic.net]
Sent: 29 August 2001 02:35
To: Chris Partridge
Cc: Ian Niles; Standard-Upper-Ontology (E-mail);
"mailto:pcassidy"@bellatlantic.net
Subject: Re: SUO: Organization
Chris --
re: your question:
Chris Partridge wrote:
> The notion of membership is slippery (and not just in these organizational
> cases). In the context of setting up companies/partnerships it seems to
mean
> owners - and hence works well with associations and co-operatives. But, as
> CYC notes, it is also used for membership of the board of directors. It
also
> seems to me that with the notion of stakeholders, employees are also some
> kind of members.
>
> Do you have some grasp of the CYC notion of member? It seems to work for
> non-commercial associations but less well for corporates.
> And why are employees not members?
>
I don't have a good grasp of how CYC treats organization
membership, beyond what you have already mentioned; however,
though the public CYC doesn't specifically include Employees
as members, it doesn't seem to exclude them, and I would think
that the broadest definition of Member of an organization
should include employees. But in the general case, any of
the membership categories could be empty. It seems
like a very heterogeneous category, and I think that there
probably should be some very general "hasMember" predicate,
since it is likely that different kinds of organization will
define membership differently. There probably
should be subcategories of membership, such as
owner, board member, employee, officer, rank-and-file member (one
not having any special role other than member). But
BoardMember will apply only to certain types of organization.
I didn't try to elaborate on membership in my previous note, since
I haven't looked into it carefully. But as a first suggestion,
I think it would be useful to have, in addition to "Membership"
and "RuleSet" as parts of an Organization, also a RoleSet, ie
a set of roles that members may play in an organization.
To conform to the TOVE structure, we would need to distinguish
roles from Positions, so we would also need a PositionSet.
Some of those positions and roles will be defined by the
RuleSet (officers), but many roles may be created more
or less formally (new job titles or tasks) by employees (e.g.
department heads). The TOVE "Organization-Role" looks somewhat
like a job description, but it can be filled by contractors.
The next level might look like this (just a fragment of what is
needed --
e.g. We would need to discuss what a Role is; a Position is more
formal,
and probably definable exclusively in relation to organizations -- it
is in the TOVE documentation, but not defined here):
(subclass RoleSet Set)
(subclass PositionSet Set)
(subclass OrganizationRoles RoleSet)
"RoleSet is a Class each instance of which is a non-empty Set,
all of whose elements are Roles"
(forall ?ROS
(=>
(instance ?ROS RoleSet)
(and
(exists ?y
(and
(element ?y ?ROS)
(instance ?y Role)))
(forall ?z
(=>
(element ?z ?ROS)
(instance ?z Role))))
))
(subclass OrganizationPositions PositionSet)
"PositionSet is a Class each instance of which is a non-empty Set,
all of whose elements are Positions"
(forall ?PS
(=>
(instance ?PS PositionSet)
(and
(exists ?y
(and
(element ?y ?PS)
(instance ?y Position)))
(forall ?z
(=>
(element ?z ?PS)
(instance ?z Position))))
))
"an organization has at least four different parts"
(disjoint should be asserted)
(properPart Membership Organization)
(properPart OrganizationRules Organization)
(properPart OrganizationRoles Organization)
(properPart OrganizationPositions Organization)
I'm not sure if every OrganizationRole and
OrganizationPosition (elements of the sets
OrganizationRoles and OrganizationPositions)
should also be elements of Membership --
that's probably too restrictive.
The general relations, trying to handle all organizations,
would have to be comprehensive. One question is whether
it would be worthwhile creating some parallel concepts
as subclasses of both OrganizationMembers and
OrganizationRole. Here is one fragment, with
"hasMember" used in the CYC sense relating members to
organizations.
(hasMember Organization OrganizationMember)
. . . but individual Organizations may want to restrict the
meaning of "member", so we would need a subclass
(subclass OrganizationMember DefinedMember)
Individuals can be idenitifed by their roles (this is one of
the linguistic alternations that can cause confusion):
(subclass OrganizationMember OrganizationOwner)
(subclass OrganizationMember Officer)
(subclass OrganizationMember Manager)
(subclass OrganizationMember Employee)
(subclass OrganizationMember BoardMember)
(subclass OrganizationMember RankAndFileMember)
But the roles themselves should be clearly distinct from
the people that fill them:
(subclass OrganizationRole OrganizationOwnerRole)
(subclass OrganizationRole OfficerRole)
(subclass OrganizationRole ManagerRole)
(subclass OrganizationRole EmployeeRole)
(subclass OrganizationRole BoardMemberRole)
(subclass OrganizationRole RankAndFileMemberRole)
. . . then, for general purposes anyone whose role is
formally part of the most general RoleSet or PositionSet
of an Organization could be called a OrganizationMember
(useful for certain types of inference), though individual
organizations would be free to define "membership" in
their own way. Programs would have to recognize
the distinction.
The roles and positions would be related to the
Collection "OrganizationMembers" by another predicate, e.g.
as in TOVE:
(filledBy OrganizationPosition OrganizationMember)
. . . also, though I don't see this in the
Fox-Barbuceanu-Gruninger paper:
(filledBy OrganizationRole OrganizationMember)
or, if we use a TOVE-like Organization-Agent
(or Organization-Entity?? They look to be the same
concept, but I'm not clear on this)
(filledBy OrganizationRole OrganizationAgent).
I don't have a good grasp of TOVE or the Enterprise Ontology;
I'm not sure if we could identify the TOVE Organization-Agent
with an OrganizationMember which is defined as a member
of a Collection in SUMO. In a quick scan of a few TOVE papers,
the only definitions I could find that might relate to the questions
of membership and roles were those in the Fox, Barbuceanu, and
Gruninger paper (http://www.eil.utoronto.ca/papers/model.html),
and in the Web pages for TOVE. For example, it appears that
in TOVE an outside contractor can fill an Organization-Role.
TOVE has an Organization-Entity concept that has contractors
as subclass. I would prefer to keep contractors separate
from members, so we might need a more general term such as
the TOVE Organization-Entity which would subsume OrganizationMember.
I would like the SUO to have a set of general concepts
that could include all that are used in existing major
ontologies. The above suggestion are just a start, and would
need supplementation or modification to be able to incorporate
other TOVE definitions smoothly.
Pat Cassidy
=============================================
Patrick Cassidy
MICRA, Inc. || (908) 561-3416
735 Belvidere Ave. || (908) 668-5252 (if no answer)
Plainfield, NJ 07062-2054 || (908) 668-5904 (fax)
internet: cassidy@micra.com
=============================================
=============================================
Patrick Cassidy
MICRA, Inc. || (908) 561-3416
735 Belvidere Ave. || (908) 668-5252 (if no answer)
Plainfield, NJ 07062-2054 || (908) 668-5904 (fax)
internet: cassidy@micra.com
=============================================