anticipation

[From Bill Powers (950507.1815 MDT)]

Bruce Abbott (950507.1855 EST) --

      This type of control system is often referred to as an ANTICIPATOR
     since the derivative control causes changes in the output forcing
     function in anticipation of an error in the immediate future.
     (Milhorn, 1966, emphasis mine)

Here is a more up-to-date quotation. Since everything that appears in
print is true, except that more recent printed comments are more true,
this statement must supercede Milhorn's.

     The common practice of calling rate feedback "anticipation" is a
     metaphorical usage and should not be given scientific credence.
     Anticipation implies knowledge of the future, but rate feedback
     depends only on knowledge of the past. (Powers, 1995, message
     950507.1815).

···

----------------------------------------------------------------------
Best,

Bill P.

[From Bob C. (980712.1212PT)]

I have been contemplating how to explain anticipation with PCT. This is how I
came across my memory questions before. I know there have been discussions of
anticipation in the past, one of which I listened in on but did not
participate in. I don't recall a PCT explanation. I know imagination and
planning are PCT syntonic words. I imagine that imagination is involved in
anticipation. I do anticipate when I am planning but also at other times. I
think of anticipation as when one event indicates that another has some
likelyhood of following. Much of my anticipation is related to my
controlled variables. Often times when I am going about my business some
event outside my control occurs which indicates that I am likely to soon be
disturbed if I don't take evasive maneuvers, or indicates an opportunity for
easier control. I tend to spend more time anticipating disturbances.

In any event, I think that anticipation involves temporal memory associations
coming into imagination. Event A serves as an associative address which
indicates that event B will likely follow. I think the ability to perceive
sequences and temporal relationships aslo likely plays a role as these are
likey more stable inductions drawn from many associative memories.

I've been trying to understand how sequences of successive events are stored
so that I can understand how one event adresses the whole sequence which
include later events which have not occurred yet when I anticipate. Memory
formation follows from the organization of the hierarchy of perceptual control
systems. So there must be perceptual systems which are continually storing
events as sequences. I often imagine sequences of successive events
spatially, as future events ahead of me, and the past behind me. This is
literally true when I'm walking. When I'm writing, the past is above and the
future below. Lets say that an animal walking at a constant speed can keep
circulating in imagination (short term memory) the last 50 yds, while also
imagining 50 yds ahead which it has just seen. Then it can just spatially
associate what is ahead with what happened behind. When the animal wants to
replay the sequence in its mind, all it has to do is imagine the spatial
associtations, and walk through the imagined space. Maybe many visual
temporal events are remembered and imagined according to some type of spatial
time line like this. I think this would be consistent with the method of loci
working so well. According to the hierarchy there are lower order perceptions
which capture images of what is happening simultaneously in space. I suppose
this would be at the level of sensations, and configurations. Sequence
systems can use these lower order systems in two ways. One in that each image
could be like a single snap shot with in a movie. Second, a bunch of snap
shots can be snapped together, like with my animal example above with events
of 50 yds behind and looking 50 yds ahead are snapped together. This visual
type representation of time might help explain why the perception of one event
can instantaneously bring up a future event skipping several events in
between, or why the future event can instantly bring up the past event.

With anticipation, event A (or sub-sequence A for more specific adressing)
serves as an associative adress to higher level sequences (longer time) which
fill in (imagine), possibilities of what might happen next. In BCP on pg. 221
address signals come from higher systems, but in the case of anticipation it
comes from below, as often the event occurs outside oneself. I assume that
the storage pathway can also serve as addressing. Is it true with associative
addressing that whenever you are addressing you are also storing and vice
versa?

However anticipation works, it helps with controling my CV's, especially when
associations are reliable. With projectile motion, if you showed me the first
part of a ball being thrown, I think I can imagine the second half of the arc
pretty well, just like if you showed me half of a new pencil and covered the
rest with your fist, I can imagine the covered half. My dog can do this to
some extent as well. After reliably throwing the ball to him for a while, I
faked with only the arm motion. He turned around and ran at leas 5 yds.
before realizing he could not find the bouncing ball to track.

Interested in hearing other's thoughts about how anticipation works,

Bob C.