[FRIAM] Why depth/thickness matters

[Eric Charles]

Glen,
Thanks for the reorientation! If you want to discuss complexity, I think an
interesting question regarding perception-action systems is how much of the
complexity has to be inside the organism, and how much of it can be
encapsulated in the larger organism-environment system. The more the
complexit is spread across the system, the more the organism can get by
with much less "mental" complexity that it might originally seem. That
tension is at the heart of Gibson vs. traditional theories, though, of
course, Gibson described the tension in different terms.

A classic example is the problem of catching a fly ball. To simplify, let
the ball be flying in a vertical plane, and let the outfielder be already
on that plane (there are very similar solutions to how to get onto the
relevant plane, so being off-plane is just a distraction). One could
imagine that the catching the ball entailed calculating a parabola-like
function, based on the start point and the speed with which the ball meets
the bat, then moving to the point where the calculation requires you to
stand. However, a much easier solution is available: Look at the ball, if
the ball is optically accelerating (i.e., moving up the visual field at an
increasing speed) step backwards, if the ball is optically decelerating
step forward, if the ball is moving at an optically constant speed, stay
where you are and put your hand in front of your head. Everything you need
to "know" what to do, is "out there" in the ambient light, and if you are a
well-designed tool, getting to the right point doesn't require modeling the
trajectory of the ball at all.

A more modern example is in locomotive robotics. Companies like Boston
Dynamics are showing that you can get basic walking movements with very
little "internal computation" if you design a system that mechanically
(through tension cords, springs, and the like) accomplish much of the
balancing and coordination. Such robots perform much better than robots who
try to handle the same types of problems in an entirely computational
"central control" fashions.

However, that doesn't necessarily speak to our ability to jettison
"representation" and replace it with dynamic-systems accounts more
generally.

For that , we would probably want to go to Tony Chemero's book, which I
mentioned earlier. In chapter 4 (summarized here
<https://psychsciencenotes.blogspot.com/2011/03/chemero-2009-chapter-4-dynamical-stance.html#more>),
Tony presents two key examples: The first is the example of the "Watts
steam governor
<http://oliverstwistarts.files.wordpress.com/2009/11/watt-regulator-yafaray-002.jpg>",
which helped stop steam engines from exploding by releasing steam. It spins
when steam goes through it, the spinning creates centerfugal force which
raises some weighted arms, which in turn open the release valve more,
keeping the internal pressure of the engine relatively constant. The second
example involved an evolutionary robotics experiment at the university of
Sussex, where allowed robots to "evolve" solutions to a problem, and then
determined how they had done so after the fact. In both cases, Tony shows
that some aspect of the system is a reasonable candidate for the label
"representation", but points out that such post-hoc labeling adds nothing
to the dynamic model.  As Andrew and Sabrina summarize in their blog:

Regarding the steam governor,
"Chemero is convinced that, according to the theory of representation from
Chapter 3, θ [angle of spinning arms] is a representation of ω [steam
pressure] and thus there is a legitimate representational account of the
governor.... Chemero is happy, however... because it is not clear the
representational story adds anything to the dynamical account. Critically,
the dynamical account must come first; you can't tell a traditional
representational story without some idea of the function of the system,
which in this case comes from the dynamical account. Given that it doesn't
add anything, you might simply wish to stop with the dynamical account and
not concern yourself with the representation that is in the system"

And regarding the evolved robots,
"Under the [old system] there is still a representational account for this
robot. The system contains visual input nodes ('representation producers')
which produce activations across intermediate nodes ('representations')
which affect the behaviour of motors via three other nodes ('representation
consumers') to produce the tracking behaviour ('adapting the system to some
part of the environment'). But Chemero describes (p.77) how this
representational gloss doesn't help - it could only be constructed after we
had the dynamical account, and the dynamical account already provides a
complete characterisation of all possible behaviours: we can use it to
predict behaviour with no reference to the representational story. Taking
the dynamical stance has 'paid off', and while it remains an ongoing task
for dynamical systems cognitive science to actually produce these types of
models, there are already numerous examples in the literature of dynamical
accounts of complex behaviour which make no reference to representation.

So, to recap: The questions for the list are 1) Where will we look for the
complexity in question? In the organism, in the environment, or in the
system that includes both? 2) Once we have a decent account of that
complexity, is anything added by inserting representation-talk in the
middle of it?








-----------
Eric P. Charles, Ph.D.
Supervisory Survey Statistician
U.S. Marine Corps
<echarles at american.edu>

On Fri, Feb 10, 2017 at 10:31 AM, ┣glen┫ <gepropella at gmail.com> wrote:

>
> On 02/10/2017 05:05 AM, Eric Charles wrote:
> > How did this all start again? Where are we going? Did I miss anything
> crucial?
>
> I started it because of the sentiment that we don't talk much about
> complexity on the list.  I think you've done a great job addressing the
> Hoffman paper in your/Holt/Gibson context Stephen appealed to.  But what
> concerns me most is that Hoffman (by virtue of games and simulation) has
> made some of the complex systems aspects of the problem explicit.  Of
> course, I'm a simulant (or "simulationist" if you must).  So I'll _always_
> throw the M&S wrench into the middle of it. 8^)  The tool is always more
> important than the use to which the tool is put.
>
> Thanks for addressing it from that context.  I'll try to comment
> constructively after others weigh in.
>
> --
> ␦glen?
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
> FRIAM-COMIC http://friam-comic.blogspot.com/ by Dr. Strangelove
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20170214/067d7115/attachment-0002.html>