Definition of "representation"

[Hyunmin Cheong]

I'm currently developing an ontology for physics-based simulation in engineering, which (I think) requires defining information content entities that represent portions of reality involved in some physical processes.

My question is on "representation". There seems to be a contradicting definition / usage.

I have read http://ceur-ws.org/Vol-1515/regular10.pdf, which states that "‘representation’ is thus more comprehensive in scope than ‘ICE,’" and defines representation is a quality.

In https://www.ncbi.nlm.nih.gov/pubmed/21143905, representation is defined as a dependent continuant (which itself does not contradict with the above definition). But in Figure 2 in the paper, representation does not fall under quality, but directly under dependent continuant (therefore, is NOT a quality).

My intuition was that representation is a generically dependent continuant (and not quality), much like an information content entity, except that the former just needs to be intended to be about something (i.e., it may refer to incorrect or non-existent things in reality).

Could someone please clarify this issue?

[Hunter, Larry]

I agree the discussion can be a bit confusing, but since your simulations all (presumably) use computers, you are safe with the IAO versions of ICE, etc. The stuff in the ICBO 2015 paper is primarily to deal with aboutness that depends on a non-continuant, most importantly on some mental process (like observing or measuring). Those are cognitive representations (MFO_0000031).

So long as your representations are dependent on a continuant (like a disk, cpu or other computer hardware), you are fine sticking with information artifacts.

With regard to the quality portion of your question, it was first (to my knowledge) introduced in the military IAO paper (http://ontology.buffalo.edu/smith/articles/STIDS-2013.pdf see especially table 2), where the idea is that the information quality is the aspect of the information artifact that bears the information, for example, the layout of pixels on a screen, or the shape of the pen marks on a paper. I would doubt you need to worry about such qualities for physics based simulation in engineering. Such qualities became a concern when trying to figure out how a (mental) process could be about something.

> My intuition was that representation is a generically dependent continuant (and not quality), much like an information content entity, except that the former just needs to be intended to be about something (i.e., it may refer to incorrect or non-existent things in reality).

An MF:representation (MFO_0000030) and an IAO:information content entity ((IAO_0000030) are both generically dependent continuants that are about a portion of reality. The difference is that MF:representations do not have to depend on continuants, they can depend on occurrents as well.

A representation in a computer, is dependent on the computer, so is safely modeled as an IAO information content entity (IAO_0000030). Depending on what kind of representation, it might fall into a defined subclass (e.g. a script is directive information entity, IAO_0000033).

Hope this helps,

Larry

[Alan Ruttenberg]

On Mon, Nov 27, 2017 at 12:18 PM, Hunter, Larry <Larry....@ucdenver.edu> wrote:

I agree the discussion can be a bit confusing, but since your simulations all (presumably) use computers, you are safe with the IAO versions of ICE, etc.  The stuff in the ICBO 2015 paper is primarily to deal with aboutness that depends on a non-continuant, most importantly on some mental process (like observing or measuring). Those are cognitive representations (MFO_0000031).

So long as your representations are dependent on a continuant (like a disk, cpu or other computer hardware), you are fine sticking with information artifacts.

With regard to the quality portion of your question, it was first (to my knowledge) introduced in the military IAO paper (http://ontology.buffalo.edu/smith/articles/STIDS-2013.pdf see especially table 2), where the idea is that the information quality is the aspect of the information artifact that bears the information, for example, the layout of pixels on a screen, or the shape of the pen marks on a paper.  I would doubt you need to worry about such qualities for physics based simulation in engineering. Such qualities became a concern when trying to figure out how a (mental) process could be about something.

> My intuition was that representation is a generically dependent continuant (and not quality), much like an information content entity, except that the former just needs to be intended to be about something (i.e., it may refer to incorrect or non-existent things in reality).

An MF:representation (MFO_0000030) and an IAO:information content entity ((IAO_0000030) are both generically dependent continuants that are about a portion of reality.  The difference is that MF:representations do not have to depend on continuants, they can depend on occurrents as well.

 

I just checked MF:representation as it was news to me that there was anything that admitted to possibly depending on an occurrent. As it turns out, MF:representation, at least formally, can't depend on an occurrent. MF:representation is a specifically dependent continuant. It is not a GDC. It is comparable with IAO:information carrier , in that its relation to an information content entity is concretization. There are two differences between MF:representation and IAO:information carrier. The first is that the MF:representation is less constrained - while IAO:information carrier is restricted to be a quality, MF:representation needs only be a specifically dependent continuant (parent of quality). The second is that in MF's formulation the concretizations/representations are individually 'about', whereas in the current IAO formulation the domain of 'is about' is information content entity (a GDC).

That said, there is a long-standing argument, which I believed I've made progress with Barry, that information can be concretized in processes, not just specifically dependent continuants. One example would be a delay line memory, where information is remembered only so long as a wave is propagating in the line. Another would be a propagating radio wave. That concretization may be in a process was always my understanding of concretization, and is apparently also the position of Roman Ingarden, from whom Barry took inspiration for the term concretization. I expect that there will be, at some point, a revision to IAO to reflect this. 

That said, Larry's advice below is what I would also say. An MF:representation doesn't capture something that is shared - it is a particular in an individual - a particular state of one instance of running of a computer. An information content entity is that which is the same across runnings of the computer/ copies in different media. 

Note that "representation" as a word is used in many ways. As with other words that have many meanings, you should be careful about evaluating a particular use, such as that in MF, against intuition. MF:representation means something specific in MF and shouldn't be understood to be the go-to term for any use of the english word "representation". 

Regards,

Alan

A representation in a computer, is dependent on the computer, so is safely modeled as an IAO information content entity (IAO_0000030).  Depending on what kind of representation, it might fall into a defined subclass (e.g. a script is directive information entity, IAO_0000033).

Hope this helps,

Larry

--
--
information-ontology@googlegroups.com
To change settings, visit
http://groups.google.com/group/information-ontology

---
You received this message because you are subscribed to the Google Groups "information-ontology" group.
To unsubscribe from this group and stop receiving emails from it, send an email to information-ontology+unsub...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.