[iDC] Embodiment is about action in context

[Usman Haque]

>I would love to learn more about Pask's "machines with underspecified
>goals".

on pask and "ill-defined goals" the best i have read is peter 
cariani's paper on attempts to rebuild a paskian electrochemical 
"ear":

http://homepage.mac.com/cariani/CarianiWebsite/PaskPaper.html

the reasoning behind 'ill-defined goals' is that if a designer 
specifies all parts of a design and specifies all behaviours that the 
constituents can have, then the embodiment will be limited by what 
the designer can predict - it is closed to novelty and can only 
respond to observables that were explicitly built into it.

if, on the other hand, an embodiment can "choose" what it senses 
(i.e. either by having "ill-defined" sensors or by determining its 
own "perceptual categories") then it moves a step closer to true 
autonomy, which would be required in an *authentically* interactive 
system. the human component of interaction then becomes crucial in 
such environments by *productively* engaging in "conversations" with 
the environment. if such an embodiment has 'ill-defined' goals then 
we humans are more able to collaborate and converge on shared goals.

this is a completely different notion of 'interaction' from that used 
in most of today's so-called interactive systems, which are premised 
on unproductive and pre-specified circular reactions.

btw, cariani's paper and several other fantastic explorations of 
pask's work can be found in his festschrift, which is unfortunately 
very difficult to find these days; if you come across a copy, grab it 
and guard it with your life...

usman