Behavior of Organisms

Phenomena
1

I don’t like any of the subcategories to the Phenomena category. It makes it difficult to know where to post. I think we only need one subcategory to Phenomena called Behavior of Organisms. I don’t know how to create a subcategory so I’m just creating this topic called Behavior of Organisms and I’ll start it off with a video of a very interesting example of behavior that turned up on Twitter – a horrible medium for having quality discussions but a great one for having dysfunctional ones. The video is here.

It’s not clear whether the cats were trained to do this or if this was their first try at getting through the obstacles. But regardless of the background – and assuming this video was not photoshopped – this is a rather fascinating example of behavior. And somewhat challenging for a control perspective since it seems like the cats are carrying out a plan of action “open loop”.

So does PCT not apply to this behavior or is their some evidence of possible controlled variables? Be sure to put on your control theory glasses before you answer!

1 Like
2

This video goes under ethology.

The comments on the video suggest that this is not consistently observed, with statements that not all cats do this. At least some clips show the posture of ‘stalking’ (body lowered, feet carefully lifted with minimal rustling or stirring around of stuff underfoot, gaze intently ahead); all the cats appear to be controlling to be at a destination behind the camera. One shows a pair of cats scoping out the (apparently) novel situation. That may be in part a social phenomenon (both will be moving through the same obstacle field and may likely control not to interfere with one another).

The ‘registration’ of footfalls calls for investigation. What would be the perceptual input? How well has ‘blind’ proprioception been investigated? Drop something under the table, take a peek to see its location, reach down and pick it up with the table surface blocking your view.

Do you dislike the subcategories under the Research category? The idea is that each subcategory of phenomena is mirrored by a subcategory of research, but there are additional research subcategories. The idea is to look for phenomena related to areas of research, and to come up with ideas how to test and research phenomena.

Your scheme of a spreadsheet database works in each subcategory. You may recall that when I proposed some phenomena outside the motor control examples with which Bill started that off, and which you have continued, I had some trouble fitting them into the pre-established columns.

3

Rick,

That’s a great video - one, to me, that PCT seems to explain well with references of getting to the other side and not touching any of the objects on the way (just as many of us do when there are goose or other animal feces on the ground).

Rich
PS: I mention goose feces because there are lots on them on the sidewalks at Fenway Park, Boston, where I often walk when visiting family near there.

4

Certainly reasonable but why ask people to have to do this classification in order to get their behavior into the database? The person who finds an interesting example of behavior may not be familiar with what these categories mean. Why not just let people put their example into this one general topic area – “Behavior of Organisms” – where we can do a preliminary PCT analysis of the behavior in terms of possible controlled variables. If it seems useful then we can classify the behaviors after we’ve got the data.

Yes, I think the fact that they are gazing ahead is particularly interesting. How do they so precisely avoid the obstacles without seeing them? Or do they see them? I think that the first order of business would be to determine whether or not the cat can see the obstacles when gazing forward. It’s possible that the obstacles can still be seen in the lower part of the cat’s peripheral visual field.

One test of this would be to train cats to run through the obstacles wearing a collar with a transparent shield below the chin. Once they have learned to do that then switch them to an equivalent collar with an very opaque shield. If it presents no problem then that’s a pretty good indication that the cat is not controlling a visual perception of the obstacles in order to avoid them.

Right. If, indeed, the cats are not seeing the obstacles then what could they be perceiving that allows the cats to avoid the obstacles so precisely while moving so swiftly? They could be controlling proprioceptive perceptions but it’s also possible that they are controlling auditory and even olfactory variables.

I do.

That’s a good point. But I don’t think the categories would help that much and they might even be a hindrance to finding relevant research. I think the best way to look for such research is to start with the phenomenon of interest itself. For example, whoever is interested in pursuing background research on this “catwalk” phenomenon should find out who made the video and go from there.

I think the latest version of that spreadsheet is here. This “Behavior of Organisms” topic would be aimed at further populating that spreadsheet. I would really like to start working on it again. I don’t remember your having problems with fitting non-motor control examples of behavior into the columns. The columns, which are labeled “Behavior, Controlled Variable(s), Reference State, Means, and Disturbances”, should work with examples of any kind of behavior. But I think it would be very helpful (to me, anyway) If you could describe an example of non-motor behavior that you had trouble fitting into my pre-established columns in the spreadsheet. Maybe we could figure out how to do it this time.

Best, Rick

5

Ah. I think I see a disparity in our perceptions.

I don’t.

Your spreadsheet is not a topic or category in Discourse. It is a Google sheet shared in Google Drive. Its purpose and function are very different from the purpose and function of topics, which is discussion. I think its relation to Discourse is best implemented as follows. Create a subcategory called CV Database (or some other 2-word name you prefer), in which the About topic contains a link to the sheet in Drive. Clicking the link always takes the user to the current state of that spreadsheet.

If someone comes up with a CV to put in the DB, great. But there is likely to be a need to discuss the addition. An added column could optionally contain a link to the topic for that discussion. They might want to open discussion of it in Discourse. There might be disagreement; refinement or elaboration might be called for. They might learn something from the discussion; they might be wrong; they might worthily challenge some unexamined assumptions; and so on.

There could be other topics in the ‘CV Database’ topic, in addition to the About topic, if there were some discussion about the database itself, its organization, usage, etc.

As you suggest, it is also possible that discussion of what seem to be control phenomena can lead to specific CVs to put in the database.

For example, we’re speculating about what cats are perceiving and controlling in that video. Perceiving first, then controlling. Just limiting the discussion to vision, cats’ eyes work somewhat differently from ours. They greatly enhance night vision at the cost of being nearsighted ( 20/100-20/200), which serves them adequately as preferentially nocturnal predators. Wider field of view and binocular field about 130º vs 120º for humans, proportionally larger (8 inch diameter human eyes would be comparable), and in my brief foray into this I still have very little idea what difference it makes to have a vertical stripe (same as dogs) rather than fovea.

With humans, as we investigate higher in the hierarchy we find differences in how perceptions that by some criteria seem the same are constructed differently. I think you will enjoy Joel Stein’s In defense of elitism for one kind of example. (I read most of one chapter in my brother-in-law’s house, the 80-year-old birthday boy in the ‘skilled nursing center’, before his youngest brother glommed onto it.)

Because this is going to continue to grow over time, with more and more topics posted, and without the subcategories, you just get a longer and longer list of disorganized topics that gets harder and harder to navigate and engage with. Discussion is liable to slip from one issue to another in the same post, and we’ll have the jumbled and often mislabeled so-called threads of CSGnet all over again.

The purposes you propose for your “one general topic area – ‘Behavior of Organisms’” are actually the purposes of the Phenomena category. We could rename Phenomena as Behavior of Organisms. Personally, I prefer Phenomena, under the maxim “phenomena first”.

A Google search did show me a couple of attempts to integrate Google Drive sheets into Discourse. As an admin, I have scant interest in taking on more support risk, and I think one link from the About topic will serve just fine.

6

I know. But it did come up in this discussion of the subcategories of the Phenomena category when you said: “Your scheme of a spreadsheet database works in each subcategory”. That’s why I asked the question you quote above: “why ask people to have to do this classification in order to get their behavior into the database”?

Far more interestingly, in the same post you said: “You may recall that when I proposed some phenomena outside the motor control examples with which Bill started that off, and which you have continued, I had some trouble fitting them into the pre-established columns”.

Given the context in which it was said – your explanation of the benefits of classifying behavior by the particular subcategories you selected for the Phenomena category – I thought you might be suggesting that these subcategories were a possible way to overcome your problem. So I asked “If you could describe an example of non-motor behavior that you had trouble fitting into my pre-established columns in the spreadsheet”. I couldn’t find an answer to that in your post so I wonder if you could describe such an example when you get a chance. It’s important to me because I think the pre-established columns in my spreadsheet – behavior, CV, reference, means, and disturbance – do cover the essential elements of all behavior, whether it’s considered to be motor or non-motor.

Best, Rick

7

I thought I posted this at the time (the end of October!) but evidently I didn’t, and I see it was because I hadn’t answered as fully as I wanted.

Yes, the column heads name the essential elements — behavior (observed phenomena), CV, reference value, means, and disturbance. Specifying these with actionable precision becomes more difficult at higher levels — actionable either for experiment or for modeling. Under the means column, it is necessary to specify these essential elements at each successively lower level of control (presumably by reference to other rows of the table). Without that, we risk mere hand-waving. Or multiple instances of the CROWD program with the interpretive labels changed.

Many such perceptions are involved in diverse instances of the Matthew Effect. (I also referred to it here).

8

The “behavior” column is NOT the observed phenomenon in PCT; it is the “named” phenomenon, like “walking”, “talking” and “playing chess”. The observed phenomena are the controlled variable (CV), the reference state or value of that variable, the means used to keep that variable in its reference state and the disturbance(s) that affect the state of that variable.

Again, you just say that this is the case. I asked for an example of a higher level controlled variable that is more difficult to specify than a lower level one. The statement above is not such an example.

I don’t believe that this is the case. We just have to see the observable outputs that are keeping the CV in the reference state. In my program control demo, for example, it is easy to specify the means by which the program is maintained in its reference state. And it’s easy to specify the program (the CV) that is being controlled as well as its reference level. All can be readily observed, as demonstrated by Powers in the Experimental Methods chapter of B:CP and by me in SLCS (The Study of Living Control Systems)

From what I can see, the “Matthew Effect” is just a statistical generalization. Of course the individuals involved are keeping all kinds of different CVs in different reference states by different means, protecting them from different disturbances. I don’t see what that has to do with the adequacy of my spreadsheet framework for analyzing the controlling done by living systems.

And by the way, the spreadsheet format we were discussing, which can be found here, is essentially the same as that described by BIll in his “Cybernetic Model for Research in Human Develoment” paper. I can’t take credit for it; it’s Bill’s idea and, as usual, a pretty good one.

Best, Rick