Model Volition or Intention

Copyright 2008 David Dodds

Is it *actually (ie verifiably) true* that all things which have a
brain, or are said to have a brain,

are or 'have' self-awareness? Or, is this just an assertion from the
'elan vital' era?

Chimpanzees are capable of understanding seeing their own image in a
mirror. Many animals are not capable of recognizing themselves in a
mirror and see their own image as being another instance of the kind
of animal they are, and often attack that image. Gorillas and
chimpanzees will use a mirror to examine their own face, this shows
self awareness. Fish and most animals simply attack any image in the
mirror that they see. One might set up cameras at the edges of lakes
and ponds so as to capture thousands of hours of animals coming to
those waters edges for a drink. And review such hours of images to see
if any of the animals show any kind of reaction or behaviour
reflecting that they caught a glimpse of their own image in the water
and had the slightest clue as to what they were seeing.

Does that little house spider in the corner sitting patiently on his
web 'have' self-awareness or shall we deny that he has a brain
instead? What about a tarantula, he has a bigger brain than that,
'is' he self-aware? A shrew? A mouse? A pekingese? (not sure it even
has a brain) Is there some scale of brain-size or -complexity that one
needs to keep in mind in making assertions about 'have'
self-awareness? Does the literature on (human) brain-pathology and -
trauma have useful information about exceptions in (adult and
otherwise) human 'is' / 'have' self-awareness? When one is "under"
anaesthesia at the hospital is one self-aware? If not where did
the self-awareness go to? Did that person's spirit go someplace else,
or did it 'stop'? We see insects brushing their antennae with their
front legs, in (what to us is) an obvious cleaning gesture,

but is the insect himself 'aware' that he is 'cleaning' his (own)
antennae (part of him'self')? Or, is the insect merely a meat robot
that when neurlogical subsystem detects that at the location of
the antennae space have 'extra (or non-original) material' on them
that a particular chorus of actions is unleashed, but which have no
more awareness (accompanying them) than the parade of figures in a
cuckoo-clock at the top of the hour?

There should be a *signal* in that we use both the term 'is' and the
term 'has' with regard to this magical thing self-awareness, and also
that we use the forms ('is') 'self-aware' and ('have') 'self
-awareness'. That is, 'to be self-aware' and 'to have self-awareness'.
('*to be* self-aware' and '*to have* self-aware*ness*'.) The former
states that self-aware is a component or part of the 'being', an
aspect of the 'being'.

'To be', the copula verb, is an interesting exception in English
language verbs. Copula verb seems to be a term which has origins in
'to create or to generate', so (human) 'being' seems to be a
generated instance of some 'thing'. Which in our case happens to be
biological.

The latter, 'to have self-awareness', suggest by that language usage
that self-aware is an attribute about ourselves. For example, we
'have' green eyes' not we 'are (form of 'is') green eyes, and we
'have' brown hair' not we 'are' ('is') brown hair' . {greeness of
eyes, brownness of hair}.

There are two separate words 'is' and 'have' in English because they
express two different things. They are not (meaning or refering to)
the same (concept) thing.

The other point to examine besides the 'elan vital' origin of
(supposedly built-in) self-awareness is what does the term
self-awareness 'mean'. If no hocus-pocus whisps of ether (like the
phlogiston theory) are evoked as explanation for what
self-awareness 'is' then perhaps a rational discussion may be made.
One thing one might do in this direction is to list things which are
and are not (given to be) attributes of self-awareness. This
would help in eliminating what self-awareness isn't, so reducing the
constellation of things to consider (that it 'is'). For human thinking
this amounts to reducing the complexity of the task.

(reducing the number of trees, so we can see the forest instead of
individual trees, which likely exceeds the magic number 7 (plus or
minus two)). {incidently, it was later found that most average
people are magic number five plus or minus two}.

new, Part Two

Is it possible for a non-human to 'have' volition or intention? Is it
possible for a non-biota to 'have' volition or intention? There are
any number of teachings, 'truths' and edicts which close one's
mind to discussion of such topics and those which include 'thinking'
and 'understanding'.

Is it possible for a machine to usefully or gainfully model volition
or intention? Instead of fretting about awarding computers with the
'elan vital' of 'having a mind' one can talk about modeling XYZ.

The values as well as the pitfalls of modeling are well known and
knowledge of both of these can allow us to scrutinize any models we
develop somewhat critically instead of taking the output purely on
faith or authoritative assertions. It seems to me this is what Alan
Turing was trying to accomplish with his Turing Test. Turing was able
to demonstrate that , via the Turing Machine concept, one could have a
rigorous description of computable equivalent. Automation and
especially programmed toys, like Asimo, AIBO, and the robot bear toy
used in Japan to comfort elderly people, have shown that (normal)
humans seek companionship to solitariness, and that movement
and (even a hint of) intellect in that movement is necessary in order
for it to be adequate.

A stuffed bear can bring comfort to a child, but it didnt seem to do
much for the lonely elderly in Japan. However there was a dramatic
change in the elderly person's interaction with a stuffed furry toy
teddy bear when it moved (by itself, via electric motors etc).
Interest disappeared quickly if the movements seemed to be random,
uncoordinated. Even if sequences of movements occurred which seemed
coordinated with each other, if they appeared exactly the same way
each time they were soon seen as 'clockwork' and 'dead'/'lifeless' and
not of further interest.

What researchers found was a) that if the robot fur-balls were
equipped with (hidden) sensors and b) when such sensor was "activated"
or 'sensed a change of some kind' by virtue of 1) physical
manipulation by an elderly person, such as simple touching, or lifting
(also notice how engaging the Wii game consoles are because of the
motion detection) or 2) by some (important) environmental
change (such as temperature), that some physical action or sequence of
physical action was made by the toy bear that seemed 'appropriate'
THEN the toy bear was taken up by the elderly person
as viable means of dealing with their loneliness/'separation'.

Apparently none of the elderly people believed that the toy bears were
'alive' in any sense that biological things are alive, none of them
mistook the toy for a living thing.

If they knew it wasnt alive what was the important difference in the
'mechanical' bears that made them acceptable or non-acceptable as
'companions' (or at least viable distractions (from loneliness)? It
seems the two biggest (most important (or 'overriding')) factors were
1) the toy physically resembled some biota that was known to be alive
generally, bears in particular.

Because it was toy bear sized it was not threatening due to size
competitive with the elderly person's size. An actual size toy grey
wolf, however delightfully furry, would not be greeted with
the same interest as a companion by an elderly person, and even less a
full sized robot bear. The quality of the 'fur' on the toy was
important. (It is my belief that an infantile memory of the hair
on "mom's" head stays with us for life, and that is why
petting'/'patting' furry animals is soothing to most people (because
it re-energizes very distant memories of safety and comfort from our
babyhood). 2) the (physical) animation of the toy bear seemed, to the
elderly person, to be in response to them or something relevant in the
local environment (like excessive room temperature).

If the bear just emitted a beep when it was touched that would not be
of interest to the elderly person even though that was clearly the
bear responding to the person's touch. If, on the other
hand, the bear moved its front legs to positions that a human infant
would to signify "up" (pick me up) if the elderly person touched the
bear in certain general areas such as its 'side' (rather than
grasping its ear), this (apparently situationally coherent motion)
would 'connect' with the elderly person. Although not entirely
rational, if the bear emitted the spoken word "up" (spoken in
Japanese) in synchrony with moving the arms there seemed to be a even
greater 'acceptance' or 'buyin' (as in credibility).

If it was reasonably possible for the elderly person to assign
'understanding' (however diminutive) to/in the bear's animation, esp
responses to physical handling of the bear such as touch and lifting
then the elderly person became somewhat 'accepting' of the bear, the
toy bear was taken up by the elderly person as viable means of dealing
with their loneliness/'separation'. Apparently the magic ingredient is
that the bear was interpreted by the elderly person as _being
responsive to her/his immediate presence_. In a sense it seems as
though the elderly person was crediting the bear with acknowledging
their existance, in the here and now. Apparently some teenagers feel
in need of this too. (The TV set cannot provide this.) Perhaps this is
why there is some appeal to elderly people for having Canaries or
Budgies around. Or cats and (generally, not large) dogs,
although the extra sappiness of Golden Retreivers makes them an exception.

What is of interest is that a known to be non-biological non-living
animal-simulacrum _appears_ to give 'intelligent', and even
somewhat-caring response to the elderly person (such as the "up"
sequence above). If the arm movement (and other movement sequences)
are programmed to have some (small excursion) random variability to
them that greatly increases the _apparent_ 'lifeness'
factor of the toy bear. In real (normal) humans we rarely do the same
motion-sequence 100% exactly the same way each time we do it. Perhaps
that's why synchronized swimming or marching soldiers is a bigger deal
than everyday life movements.

If we use the terms volition and intent in the general way that we do
when we use these terms to talk about ourselves (humans) its hard to
point out where or how either of these things are present in these
bears above. If one has some decent ACTUAL knowledge about programming
computers and about robotic devices (as opposed to Godlike (but wrong)
opinion) then one is hardpressed to identify any volition and intent
in those bears (in the way when we use these terms to talk about
ourselves (humans)). Yet it is the very (apparent) perception of
something like these occuring in the machine bears that makes them
'accepted'. It could be said that when there is an interest in a human
to do so they can find apparent cues to attribute mental processing,
such as 'understanding' (of touch, lifting, etc) in events where it
(apparently) is not actually present.

Certain sounds made by a dog are heard as a "groan" and we infer that
the dog would have 'rolled his eyes' too if he could. We infer from
the sound that the dog is having a mental event similar to our "oh
brother..".

When the programming of the bears is such that the programming takes
care to use just-previous action sequences as context (for following
sequences) this improves the otherwise 'mechanical-ness' (appearance)
of the bear's 'behaviour' so that it seems less like a clockwork and
more like a biologically 'organically organized' entity. As such
machines (as these bears) have incorporated more subsystems like that
of the MIT KISMET robot one will see people more and more drawn in
by such machines to participate or engage with them in social
interaction. This is reality at present.

Some possible futures may be seen in the (dark) movie "AI". We have
already seen that actual positive social benefit can occur when humans
(such as elderly) are willing to suspend certain notions and give the
benefit of the doubt to (dead) mechanical robots, vis-a-vis some
'understanding' and 'intelligence' (however little it might be). Small
children dont seem to have grasped the difference between
animated/moving things and living things, once cars are self-
driving (already done experimentally) and have extensive voice
feedback (like some GPS systems) even adults may wonder if the thing
they're riding in is completely dead.

This discussion leads us into future discussion of planning-software,
and debate about how sophisticated a plan can be attained by a planner
(program) where 'nobody is home' (ie the program is running as a
complex algorithmic system, rather than an agent which 'understands'
both the problem and its solution (behaviour)).