Inverse of is_about?

[Janna Hastings]

Hello,

quick question. What is the relationship to use for the inverse of is_about? That is, I have some (for example) molecular entity, and I have some information content entity which is about the molecular entity. The information is about the molecule, but what is the molecule with respect to the information? is_described_by? is_subject_of?

Help appreciated!

Best wishes,
Janna

[Alan Ruttenberg]

On Fri, Oct 16, 2009 at 9:08 AM, Janna Hastings
<janna.h...@gmail.com> wrote:
> Hello,
>
> quick question. What is the relationship to use for the inverse of is_about?
> That is, I have some (for example) molecular entity, and I have some
> information content entity which is about the molecular entity. The
> information is about the molecule, but what is the molecule with respect to
> the information? is_described_by? is_subject_of?

I've been trying to avoid having this relation because it reinforces
the idea that there is a bona fide relation originating from the thing
and towards the information, and it seems like a classic error to say
e.g. that a name is a property of the thing it names.

I have several thoughts:

a) OWL 2 allows one to use inv(is_about) in class expressions. Use that.
b) Since sparql is relational in its form, you don't need the inverse
relation to do a query. Just reverse the order of arguments.
c) Define a relation but have some marking to indicate it's status as
a "junk relation".

-Alan

[Janna Hastings]

Hello,

On Fri, Oct 16, 2009 at 2:13 PM, Alan Ruttenberg <alanrut...@gmail.com> wrote:

> quick question. What is the relationship to use for the inverse of is_about?
> That is, I have some (for example) molecular entity, and I have some
> information content entity which is about the molecular entity. The
> information is about the molecule, but what is the molecule with respect to
> the information? is_described_by? is_subject_of?

I've been trying to avoid having this relation because it reinforces
the idea that there is a bona fide relation originating from the thing
and towards the information, and it seems like a classic error to say
e.g. that a name is a property of the thing it names.

Gotcha. This makes sense.
 

I have several thoughts:

a) OWL 2 allows one to use inv(is_about) in class expressions. Use that.
b) Since sparql is relational in its form, you don't need the inverse
relation to do a query. Just reverse the order of arguments.
c) Define a relation but have some marking to indicate it's status as
a "junk relation".

I am happy to always model aboutness in the ontology in the direction from the information to the molecule (for example). I would merely like to be able to speak in a natural way in the inverse, since it so often is the case that molecules are only synthesised to match graphs (for example) that have already been in existence for a while. The molecules are synthesised to match a kind of specification, but the specification is not really about the molecule until the molecule has been synthesised.

So I think option (c) above works for this use case.

Thanks,
Janna

[Colin Batchelor]

On Oct 16, 2:26 pm, Janna Hastings <janna.hasti...@gmail.com> wrote:

> I am happy to always model aboutness in the ontology in the direction from
> the information to the molecule (for example). I would merely like to be

> able to *speak *in a natural way in the inverse, since it so often is the

> case that molecules are only synthesised to match graphs (for example) that
> have already been in existence for a while. The molecules are synthesised to
> match a kind of specification, but the specification is not really about the
> molecule until the molecule has been synthesised.

I think a more process-oriented, OBIish way of doing this would be
better: your molecule is the output of a (family of) planned processes
which have different effective specifications (add 50 mg of compound 3
to 0.1 M HCl etc.) but the same ineffective specification (the graph).

Best wishes,
Colin.

[Larry Hunter]

See comments at the end...

In this case, it seems to me that the molecular graph is more than
"is_about" the molecule. Here, I think it sounds like a specification
of a realizable, like any other bit of a plan. The link from a
realizable entity to the information about how to realize it is
"has_specification", which I think is better than making up a new
"junk" relation. Would that cover all of your use cases?

> So I think option (c) above works for this use case.

I'd prefer not to go in this direction if we don't absolutely have to.

Larry

[Goldstein, Adam]

Comments @ end.

-----Original Message-----
From: informatio...@googlegroups.com on behalf of Larry Hunter
Sent: Mon 19-Oct-09 11:40
To: Janna Hastings
Cc: Alan Ruttenberg; informatio...@googlegroups.com
Subject: Re: Inverse of is_about?

On Oct 16, 2009, at 7:26 AM, Janna Hastings wrote:

> Hello,
>
> On Fri, Oct 16, 2009 at 2:13 PM, Alan Ruttenberg <alanrut...@gmail.com
> > wrote:
> > quick question. What is the relationship to use for the inverse of
> is_about?
> > That is, I have some (for example) molecular entity, and I have some
> > information content entity which is about the molecular entity. The
> > information is about the molecule, but what is the molecule with
> respect to
> > the information? is_described_by? is_subject_of?
>
> I've been trying to avoid having this relation because it reinforces
> the idea that there is a bona fide relation originating from the thing
> and towards the information, and it seems like a classic error to say
> e.g. that a name is a property of the thing it names.
>

There is another use case i which I'm interested---describe objects of study, etc., in terms of the works about them---someone starts with "natural selection" eg and can see what works there are which are about that process. This really does seem like a relationship of "inverse aboutness." I don't think we want to say that information content of a work, and its appearance in the work, is the result of a process or is a realizeable, though of course it is; but we can look directly at the content in this case.

=================================
Adam M. Goldstein PhD MSLIS
Assistant Professor of Philosophy
Iona College
--
agold...@iona.edu
http://www.iona.edu/faculty/agoldstein/
(914) 637-2717
Iona College
Department of Philosophy
715 North Avenue
New Rochelle, NY 10801

[Goldstein, Adam]

-----Original Message-----
From: informatio...@googlegroups.com on behalf of Larry Hunter
Sent: Mon 19-Oct-09 11:40
To: Janna Hastings
Cc: Alan Ruttenberg; informatio...@googlegroups.com
Subject: Re: Inverse of is_about?

On Oct 16, 2009, at 7:26 AM, Janna Hastings wrote:

> Hello,
>
> On Fri, Oct 16, 2009 at 2:13 PM, Alan Ruttenberg <alanrut...@gmail.com
> > wrote:
> > quick question. What is the relationship to use for the inverse of
> is_about?
> > That is, I have some (for example) molecular entity, and I have some
> > information content entity which is about the molecular entity. The
> > information is about the molecule, but what is the molecule with
> respect to
> > the information? is_described_by? is_subject_of?
>
> I've been trying to avoid having this relation because it reinforces
> the idea that there is a bona fide relation originating from the thing
> and towards the information, and it seems like a classic error to say
> e.g. that a name is a property of the thing it names.
>

I caught the paragraph immediately above as I sent my last reply. I don't think I am making the classic error? I think that the "is a subject of" is not a property of the object, but is a property of the entire complex consisting of the object and the individuals who have thought about/talked about/written about the object and the institutions and mechanisms of realizing the information. Maybe that is part of the idea of calling a graph a realizeable (mentioned by Larry Hunter elsewhere in this thread).

Adam