[topicmapmail] Re: How to best model relations with mode (assertive, modal, intention, negation) with topic maps?

[Alexander Sigel]

Dear topic mappers,

both Jan Algermissen and Dmitry Bogachev suggested using scope to model
assertations being true only in a certain interpretation (see below).
This is fine (see e.g. TMDM definitions and LTM capabilities below).

I had considered scoping associations in the first place, but had two
issues with it:

1. It can be said with equally valid justification that e.g. negation
should be modelled using:
   * scope, because it is a knowledge base partition (context, view,
interpretation)
   * reification, because it is a statement about a third-party
statement (statement, which in the context of someone else is just a
claim, or even not accepted).

   I do not know fully which are advantages and disadvantages of either
modelling choice.
   For simplicity, I now lean rather towards scope.

2. I somehow got quite confused getting the scope setting right,
   when mixing it with the common but somehow non-standard trick for
scoping names by role types to get the association name nicely
displayed.
   Maybe someone can help me out here?

Here, I use LTM for modelling and the Omnigator for displaying.

To express that the causality relation between two given topics is true
only in modal mode, it seems correct to model:
r2_4(t39 : rt_cause, t41 : rt_caused) / mode_modal

However, I can't get Omnigator to display the name I would like to have
displayed.
It does not matter if I use plain ANDed scope (role type AND mode in
scope) or (as here) variant scope for modes within role type scope.

When looking from the side of the role player type "cause",
in assertive mode the relationship name should be "IS the cause", and in
modal mode "MAY be the cause of".

Although Omnigator displays the relationship name in the scope of the
role player type nicely, it does not display the mode I wish it to
display.

Filtering along axes does not really help:
* Filtering variant name context results in ALL assertions (even not
modal ones) shown with the modally scoped name. Not ok, since only
modally scoped assertions should be shown with the modal name.
* Filtering association context for modal mode results in only the
modally scoped assertions shown. Ok, but no name change.
* Filtering both by variant name and association context displays the
desired name and association. XXX

So, is there a way to cleanly combine using scope for naming
associations according to the view of the association role types, with
this mode scope?
How would one do this?

My example:
-----------
/* association modes */
[mode_assertive : tt_mode = "assertive mode"]
[mode_modal : tt_mode = "modal mode"]
[mode_intentional : tt_mode = "intentional mode"]
[mode_negative : tt_mode = "negative mode"]

/* association role types */
[rt_cause : ICLO_role_type = "cause"]
[rt_caused : ICLO_role_type = "caused"]

/* association type */
[r2_4 : ICLO_relation = "cause-caused"
			    = "causality"

			    = "is cause of" / rt_cause
					    ("IS cause of" /
mode_assertive)
                                  ("MAY be the cause of" / mode_modal)
					    ("is CLAIMED to be cause of"
/ mode_intentional)
					    ("is NOT cause of" /
mode_negative)
					  
                      = "is caused by" / rt_caused
                                  ("IS caused by" / mode_assertive)
					    ("MAY be caused by" /
mode_modal)
					    ("is CLAIMED to be caused
by" / mode_intentional)
			       	    ("is NOT caused by" / mode_negative)
]

 
/* ----------------------------------------------- */
/* make statements */  
/* ----------------------------------------------- */

/* topics */
[t39 = "deep sea earthquakes"]
[t40 = "tsunamis"]
[t41 = "tsunami catastrophes"]

/* statements */
r2_4(t39 : rt_cause, t40 : rt_caused) / mode_assertive
r2_4(t39 : rt_cause, t41 : rt_caused) / mode_modal

The first statement says of itself that it is true (factual), the
second, that it is only possibly true.

Best regards
alex
---
Alexander Sigel, M.A., Researcher in Semantic Knowledge Networking
Univ. of Cologne, Dept. of Information Systems & Information Management
http://www.wim.uni-koeln.de/19.0.html
sigel at wim.uni-koeln.de, +49 221 470-5322

------
 
[Jan]
> I guess that is *the* use case for scope, isn't it?
> What is the reason scope is *not* your primary choice??

[Dmitry]
I would use scope in this case. Scope allows to represent different
worlds in the same upper context and even to do inference for diffrent
worlds and not to mix them.
It is also possible to "lift" assertions from one world to another and
compare worlds.
I often use these context dimensions:
* truth: false, default false, unknown, default true, true
* domain: virtual worlds, content of resources, different areas of
knowledge, enterprise applications
* opinion: person or organization
* time: interval in time
* language: (for names only)

TMDM
----
3.12 scope
context within which a statement is valid 

5.7 association items
All associations have a scope, which defines the context in which the
relationship represented by the association is considered valid. The
scope also applies to the assignment of the roles to the topics playing
them; that is, the scope defines the context in which the topics can be
said to play the roles in the association.

LTM
---
2.2 defining associations
Associations can also be scoped, as with base names, by appending a
slash followed by the IDs of the scoping topics, separated by
whitespace.