Inability to Represent Stasis and Change of Location as valid Processes with Process Profiles in CCO

[rmavrovich]

GitHub Issue #1:

@giacomodecolle and I wrote this together after discussing how to represent stasis and change of location in CCO. @FedeDon was also involved in discussing this on Slack.
In BFO/CCO, location is defined as a relation: “occupies spatial region” between a material entity and a spatial region. A stasis location should be a process where an object remains in the same spatial region for its duration. A change of location process should involve different spatial regions/sites at different time points.

Change:

The definition of “Change” in BFO/CCO defines it as:

-Increase/decrease in intensity of dependent entities.
-Gaining new dependent entities.
-Ceasing to bear dependent entities.

Issue: It excludes “located in” or “occupies spatial region” relations.

Movement and Stasis:

CCO defines "Motion” as: “A Natural Process in which a Continuant changes its Location or Spatial Orientation over some Temporal Interval.”

Issue: "location" is undefined. A reasonable interpretation is that if x moves then x is located at y at t1 and it is not located at y at t2, but again this is unclear in the definition. Yet if "change" is spelled in terms of increase/decrease/gain/loss of an SDC or GDC, then it's not clear how a continuant can change its location. A possible solution would be to add "location" as a relational quality.

CCO defines Stasis as: “A Process in which one or more Independent Continuants endure in an unchanging condition.”

Issue: If “condition” refers to specifically dependent continuants, then unchanging location is excluded. Consequently, we can't define the stasis of location or related process profiles.

Process Profiles:

Process Profile definition: “An occurrent that is an occurrent part of some process by virtue of the rate, or pattern, or amplitude of change in an attribute of one or more participants of said process.”

Issue: Movement and Act of Motion aren’t classified as “Changes” (per the narrow interpretation of the definition) and they also seem unable to have process profiles, given that process profiles can only be profiles of some type of change. Take as an example the class "Velocity", which is a process profile or "Motion". What is the change in virtue of which velocity is a process profile? Not a change in location, since that is not proper change according to the CCO definition of change.

[CarterBeauBenson]

Ray,

Importantly, "located in" is the object property you would want to use for "location" since it connects independent continuants (and not spatial regions) to independent continuants (and not spatial regions). We can see "artifact location" is the site that is the location of an artifact. We could make a defined class that is a sibling of "artifact location" that is "person location" that is the site that is the location of a person.

Per previous issues about representing distance as a relational quality, you could represent a lack of movement as a stasis of a relational quality, since the distance between me and the door will be unchanging between me and a door while I do not move. But every time I rock in my chair, I break the stasis. Maybe someone could argue that location is a relational quality of some sort like distance, but it is currently unclear to me that location change is a change in an SDC like being fired from a job.

That said, it is a little bit odd for me to think of a change in location as something you'd want to understand as stasis. Stasis should probably be reserved to talk about things like an unchanging citizen role or something like that. Do not get bogged down in how we say things like "I changed my location" and that it cannot be directly translated into CCO using the "change" class. They are just labels. It could just as easily be called "change in dependent continuant" and "act of motion" could be called "change in location".

Carter

[giacomodecolle]

Carter, if we are using "located in" over "occupies spatial region", take the example of my sitting in my chair while the Earth moves. I haven't changed my "located in" but I have changed my "occupies spatial region". Would this not be a change of location in some sense that we may want to capture?

Besides that, let's say that "Motion" is equivalent to "change of location" where location is understood as changing a relational quality between me and a site. Then I can also have "location stasis" as me not changing the relational quality of location. This seems to me to be helpful insofar we want to track locations of objects, say when they started changing location, etc.

But if this is true, then should "Motion" and "Stasis of location" (or whatever you want to call this) not be under "Change" and "Stasis" respectively? If so, would Motion go under Decrease/Gain of dependent continuant?

[CarterBeauBenson]

A lot to unpack here. It is unclear to me why would ever want to use Spatial Regions. Everything we care about can be cashed out in Sites. Processes occur at sites. MEs are at sites. Sites can have qualities, SRs cannot. The way you stated it, your movement through the universe is constant. That seems to neither be change nor stasis in the CCO sense.

"Act of location change" is a subclass of "act of motion", so its not quite equivalent. But all acts of location change are also acts of location.

You could model that, but why? I am not seeing what it would gain to create an additional group of triples for when someone moves that could already be cashed out through ICEs and acts.

[giacomodecolle]

"Act of Motion" and "Act of Location Change" are processes where an agent makes something move. I understand this is meant to be used for things like me lifting a box.

Are you suggesting that I should just have ICEs that say that I occupy site 1 at t1 and occupy site 2 at t2 for change, and that I occupy site 1 at t1 and occupy site 1 at t2 for location stasis? But how are you going to specify that temporal element? I see stasis processes to be type of things I would use to add time information through the "occupies temporal region" relation.

[CarterBeauBenson]

The key understanding here is that your site does not change. The site of Giacomo's human organism is the same regardless of whether he is in the US or Italy. The site that your site overlaps with does change from one site of a country to another.

Now think about data. We have coordinates that are measurements of your location, and that is what shows up in the data. At t1, there is a measurement of your site that is the output of some GPS process from your phone. At t2, there is a measurement of your site that is the output of some second GPS process from your phone. Then we draw bounding boxes around countries, specified by the lat an lon of the borders. We then compare to see if your lat and lon falls within the site inside of the bounding box.

This same site and bounding box method works if you change your location from your office to Fede's office.

[oliviahobai]

Given the messages here are about clarifying existing definitions & different modeling tactics, rather than changing the CCO, with everybody's assent I'm going to close this issue and ask that people continue with a discussion thread instead.