home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!auvm!BEN.DCIEM.DND.CA!MMT
- Message-ID: <9208271825.AA25558@chroma.dciem.dnd.ca>
- Newsgroups: bit.listserv.csg-l
- Date: Thu, 27 Aug 1992 14:25:07 EDT
- Sender: "Control Systems Group Network (CSGnet)" <CSG-L@UIUCVMD.BITNET>
- From: mmt@BEN.DCIEM.DND.CA
- Subject: Re: Misc Replies
- Lines: 78
-
- [Martin Taylor 920827 14:00]
- (Rick Marken 920827 10:00)
-
- I'm lost. I don't understand what it is that Rick doesn't understand so I
- can't answer his questions. Rick seems to be saying that there are no more
- sensory degrees of freedom than motor degrees of freedom, and that all
- ECSs can always be controlling. That, at least, is what I think is the case
- in the models I have seen simulated or discussed. But it can't be true,
- and I'm sure Rick knows it can't be true, any more than it can be true that
- we behave in a strict S-R manner. So, since I believe Rick to be reasonable,
- there must be some intent behind his questions that totally escapes me.
-
- =============quotes follow.
-
- All I'm asking for is some data that would motivate the need
- to revise the existing model (which handles many simple cases of control). I
- would suggest a very simple experiment; just so I could get the idea of what
- you are talking about. tI scould probably just be a simple extension of a
- multidimensional tracking task (which we know the model can handle). Just
- create a simple change in the experiment that produces consistent results that
- pose a problem for the existing model.
-
- >There really isn't any experimental demonstration that I can think of,
- >because
-
- If you really can't do an experiment that tests your proposed revision of the
- model then I can't see that it's such an important revision; since the model
- accounts for the phenomena we've observed so far I can't see revising the
- model to help it explain a phenomenon that we will never be able to
- observe.
-
- ============a sort of response? follows:
- Has the model accounted for all the phenomena of human life observed so far?
- I thought that I mentioned a couple of phenomena that seem not to be covered
- by the N-in N-out model: the flick of the eyes to an unexpected movement,
- the turn of the head toward an unexpected noise. These seem to involve a
- K-in N-out model, where K>>N. This seems crystal clear, and I don't suppose
- that's where the problem lies.
-
- All I am doing is simply pointing out a fact that must be true in the real
- world (as opposed to the world of constructed models acting in trivial
- environments). It's as much a fact as that disturbances sometimes affect
- CEVs that correspond to percepts. The latter fact leads to PCT in general,
- the df discrepancy leads to what I have called "alerting systems" within PCT.
-
- Perhaps there's a misunderstanding about the difference between being able to
- observe a phenomenon and being able to do experiments on it. Rick, is that
- where the problem lies? Astronomers can observe but not experiment. They
- can model their phenomena using the results of experiments done by physicists
- and chemists. My claim was not that there are no observable phenomena
- attributable to alerting systems, but that we couldn't undo evolution, and
- wouldn't ever find an active living system without them; and if we constructed
- a simulation, its construction would be whatever the designer decided. If it
- failed to respond to an alerting condition, the reason would be in the design.
-
- And I did ask if anyone else could think of an experimental situation that
- might tease out the mechanisms of alerting (as opposed to its existence).
- Maybe there could be experiments in the area of high workload stress, but
- they would probably be very complex, and I can't see how they would work.
-
- Finally:
- >>I doubt you could ever find a phenomenon that can be described by only
- >>one organization.
- >
- >True. But a collection of observations can usually only be explained just one
- >of the models.
-
- When you are comparing the relative merits of a fixed set of models, any
- observations can change their relative likelihoods. But there is always
- an infinity of other models standing in the wings awaiting their turn to
- be tested.
-
- I have a fundamental proposition about science that I have seen no reason
- to change since before my grad school days:
-
- "Whatever your theory, it is either imprecise or wrong."
-
- Martin
-