Relationships
dgillm...@gmail.com
Dear Colleagues,
I have a question about relationships, which has come up in some work I doing with a colleague to create a framework for the information (meanings, inferences) contained in qualitative data. Qualitative data are videos of interviews, images of artwork, focus group transcripts, etc. They require a researcher to analyze the contents of a file to uncover what they are about. Part of that “aboutness” includes relationships between things found in the files and what those relationships might mean.
Two kinds of relations, time and space proximity, are objective in the sense the closeness can be measured without caring about a meaning. Importantly, the same measurement of closeness may not end up being meaningful in the same way each time.
The question is whether there are other relations that are objective in the same way. For example, a parent/child relationship as seen in some video might be implied, but I don’t see any way that can be measured or in any other way objectively determined just from observing (not analyzing) the contents of a file. So, we know there are relations that don’t qualify as an answer to the question. Maybe an asynchronous sequential relation counts. Are there others?
This question might have been discussed earlier in the life of this email forum, and if so please direct me to the thread. In any case, I am eager to hear your opinions.
Thanks.
Regards,
Dan Gillman
PS There are purposeful differences between my usage of measure and relation on the one hand and measurement and relationship on the other. The difference in both cases is type/instance.
Alex Shkotin
Dear Dan,
To better understand your question, I'll describe the topic from the perspective of axiomatic theories and their models.
Take, for example, Hilbert's axiomatics for Euclidean geometry.
There are only two primary relations there: lies on and betweenness. All other relations will be defined through them.
Now, if I look at two non-intersecting triangles, one of which is inside the other, there is a relation "within" between them, which will be defined in the theory.
Is this relation "within" objective for you?
And what is "an asynchronous sequential relation"? Can you give an example?
If we look at "videos of interviews, images of artwork, focus group transcripts, etc." as a model for one or another theory we get one or another description of this particular "qualitative data". Or maybe you are talking about phenomenological relations?
But in this case, the parent-child relationship can only be determined by tracing it back in time to the moment of birth, and even then only for the mother.
Most relations have definitions which are very sophisticated.
We may take OBO Foundry Relation Ontology (AIed:https://share.google/aimode/QoC38rEHUnSYM1mAa) to begin with phenomenological relations.
A very interesting topic.
Regards,
Alex
--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/017901dcf1d5%2446d5ff70%24d481fe50%24%40gmail.com.
dgillm...@gmail.com
Dear Alex,
Thank you for your interest.
I think I should describe an analytic scenario involving qualitative data to try to anchor my question. Imagine looking at a video of a family riding in a car on a rainy day. The family consists of the mother, the father, and a pre-verbal one year old child. The child is riding in a car seat facing forward. In the video you can see each member of the family and the windshield of the car. The rain is sporadic, so the driver (the father in this case) sometimes has to turn the windshield wipers on, and sometimes they need to be off. The researcher is analyzing how parents treat their children when they get demanding.
And by the way, ignore the fact there are parent/child relationships among the participants. This is not germane.
After watching the video several times, the researcher notices the child starts to cry each time the wipers are shut off and stops when the wipers are turned back on. This cycles. Each interval of time where the wipers are switched on/off and the child reacts can be located in the video, and the length of the time interval can be measured. This is what I mean by objective. It is up to the researcher to decide if the coincidence (wipers off / child cries) is significant, and this might depend on how quickly and regularly the child reacts. So, the measure of the interval is a clue for the researcher. In any case there is a time-dependent relationship between what the wipers and child are doing.
This example applies to time, and it shouldn’t be hard to find examples applicable to location, which are also objective in the same way. Are there other kinds of relations besides time and location that are objective?
I hope this is clearer. At this stage, I am not interested in finding a fundamental set of relations. I am interested in knowing if there are additional relations that have “objective” as a property in the sense described here.
Thanks.
Regards,
Dan
To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxROTYHN90Td7fE88-GQa_Oh-Z%3D0ztSxeL15uJyoa8UNFmLg%40mail.gmail.com.
Alex Shkotin
Dan,
Your example is quite complex, as the video description itself contains several hypotheses that presuppose a theory. For example, the correlation between rain intensity and the driver turning on the windshield wipers is the same as the correlation between turning on the windshield wipers and a child's reaction.
Correlation has a long way to go in an objective relation. That's why I wrote about the theory of events that the researcher spontaneously adheres to. Without a theory, the only hypothesis they can put forward is that there must be some connection between the windshield wiper movements and the child's reaction.
In your example, the relationship between the windshield wiper movements and the child's behavior is unknown. But we were talking about an example of an objective relationship.
By the way, there's a website with tons of examples of unexplained correlations.
Is it correct to say that you call the relationships obtained from measurement results objective?
For example, the driver's weight is greater than the child's weight.
Then any relationship whose definition includes only measurable characteristics will be objective.
Science and technology, then, are full of objective relationships.
And it would be good if you provided "examples applicable to location" for clarity. One for objective and one for non-objective.
Your example is good as an example from Higher Order Logic: There is some relationship between the child's behavior and the movement of the windshield wipers.
Alex
To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/007f01dcf2cc%244a5b4d60%24df11e820%24%40gmail.com.
dgillm...@gmail.com
Alex,
I think you are misunderstanding the situation. The researcher might have a general question she wishes to investigate. She will locate files that contain brief descriptions of the content and select ones that seem relevant. There is no or little information about the details of what she might find.
Further, consider the fact in my example she is viewing a video. This is a rendering of a recording of some entities and event. There is no guarantee all the entities and events in the original are discernable in the video, and she knows neither what was in the original nor the limits of the recording. All she can see is what’s there.
No matter what might be left out, there are entities and events that are visible. Any researcher can see them. This is what I mean by objective. And these entities and events exist in the video independent of their meanings. For example, 2 researchers might infer different meanings to the same event.
Going frame by frame through the video, the researcher can identify segments that are of interest. Possibly in several of those segments some events are visible (wipers go off; child starts crying). The question is whether they are related. One factor might be the time interval between them. This can be measured simply by viewing the video. This is independent of what they might mean.
Your weight example requires making an estimation of those weights based on how each person looks in the video. That requires inferences. Two researchers might estimate the weights of people and come up with the opposite relationship. I am not interested in inferences here.
Maybe you will say seeing the child cry is an inference. OK, then let’s say we see the child change his facial expression with tears coming down his cheeks.
Let’s call this closeness time proximity. Location proximity can similarly be measured. Are there other relations that can be measured in this way? And realize, the measurement I am referring to is made against the video, not about the entities and events seen in the video.
Yours,
To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxROTO7rXFFnhDkUK2%3D7EqCDbgN92Gs%2B5LOco5jTAAcG-e8w%40mail.gmail.com.
Alex Shkotin
Dan
In the weight example, I meant that it's measured and known to the people in the picture. It can be replaced by a visible quantity: a tall person drives a car, a small person cries from time to time. Relationships of volume, area, and length are measurable and objective.
I truly, sincerely want to understand which relationships you consider objective and which are not.
And that's why I asked you a question to help me understand:
`And it would be good if you provided "examples applicable to location" for clarity. One for objective and one for non-objective.`
However, it seems we've gotten stuck. You ask, "Are there other relationships that can be measured this way?"
Unfortunately, I don't understand which relationship we measured and which way.
Yours,
Alex
ryo...@wwweb.org
Hi Dan,
Re/ Branching-from: “Imagine looking at a video of a family riding in a car on a rainy day. “
This recorded presentation may be of interest:
Presentation: Metadata that fosters collaboration & control — learn from keywording expert Clemency Wright
Date that presentation was held on: Thursday, May 21 at 2:00PM (BST)
Recording available by registering here: https://www.fotoware.com/resources/webinars-events/metadata-that-fosters-collaboration-control-with-clemency-wright
You should get a confirmation email, with a personal link to the recording.
(mine got caught in junk, so not sure how quick it was to arrive).
In (my) summary, the presentation/ recording...
- Talks about the field of Digital Access Management (DAM), for photos, videos.
- Discusses having Controlled Vocabularies, to aid access to those digital materials.
- An audience Practical Workshop, towards the end (time stamp @34:23), with three sample photos, asking the audience to describe each, with a myriad range of results.
- Attendees also received a follow-on pack, with useful links, like:
- DAM taxonomies & controlled vocabularies: A practical guide https://www.fotoware.com/blog/taxonomies-controlled-vocabulary-in-dam
- Content Authenticity: How to protect trust in the digital age https://www.fotoware.com/blog/content-authenticity
Seems like a good snapshot of current thinking in picture and video DAM practice.
Also, an illustration of the value and use of Controlled Vocabularies for digital media.
In answer to your question: "The question is whether there are other relations that are objective in the same way",
For me, the practice would suggest that...
- There are myriad ways of describing an object. The subjectivity of the audience-involved Practical Workshop, in the presentation exercise, was quite illuminating.
- Some ways may be called subjective, and some objective, but in many views, even objectivity (eg. space, time) is subjective/ relative.
- An answer may be to just agree what frames you are using, and hence, what Controlled Vocabularies could usefully apply, in your main situations.
=> The worlds of (i) photo and video data management, and (ii) DAM, are already quite established, with many useful frames, standards, practices etc available to use and build on.
Hope it is of interest.
My apologies if this conversation has moved on. I took some time to contact Fotoware (the organisers of the presentation for a public link to share).
Regards,
Rob Young
(not associated with the presentation in any way, except as an interested viewer of the event/ topic)
To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/007f01dcf2cc%244a5b4d60%24df11e820%24%40gmail.com.
