Stathis,
I am answering all the mail in time order. I can see below you are making
some progress! This is cool.
> Colin Hales writes:
>> >> So, I have my zombie scientist and my human scientist and
>> >> I ask them to do science on exquisite novelty. What happens?
>> >> The novelty is invisible to the zombie, who has the internal
>> >> life of a dreamless sleep. The reason it is invisible is
>> >> because there is no phenomenal consciousness. The zombie
>> >> has only sensory data to use to do science. There are an
>> >> infinite number of ways that same sensory data could arrive
>> >> from an infinity of external natural world situtations.
>> >> The sensory data is ambiguous - it's all the same - action
>> >> potential pulse trains traveling from sensors to brain.
>> Stathis:
>> > All I have to work on is sensory data also.
>> No you don't! You have an entire separate set of
>> perceptual/experiential fields constructed from sensory feeds.
>> The fact of this is proven - think
>> of hallucination. When the senory data gets overidden
>> by the internal imagery (schizophrenia). Sensing is NOT
>> our perceptions. It is these latter phenomenal fields
>> that you consciously work from as a scientist. Not the
>> sensory feeds.
>> This seems to be a recurring misunderstanding or something
>> people seem to be struggling with. It feels like its coming
>> from your senses but it's all generated inside your head.
>
> OK, I'll revise my claim: all I have to work with is
> perceptions which I assume are coming from sense data which
> I assume is
> coming from the real world impinging on my sense organs.
> The same is true of a machine which receives environmental
> input and processes it. At the processing stage, this is
> the equivalent of perception. The processor assumes that
> the information it is processing originates from sensors which
> are responding to real world stimuli, but it has no way of
> knowing if the data actually arose from spontaneous
> or externally induced activity at any point from the sensors,
> transducers, conductors, or components of the processor itself:
> whether they are hallucinations, in fact. There might be some
> clue that it is not a legitimate sensory feed, but if the
> halllucination is perfect it is by definition
> impossible to detect.
>
By George, you're getting it!
>> Whatever 'reality' is, it is regular/persistent,
>> repeatable/stable enough to do science on it via
>> our phenomenality and come
>> up with laws that seem to characterise how it will appear
>> to us in our phenomenality.
>
> You could say: my perceptions are
> regular/persistent/repeatable/stable enough to assume an
> external reality generating them and to do science on. And if
> a machine's central processor's perceptions are similarly
> regular/persistent/, repeatable/stable, it could also do
> science on them. The point is, neither I nor
> the machine has any magical knowledge of an external world.
> All we have is regularities in perceptions, which we assume
> to be originating from the external world because that's
> a good model which stands up no matter what we throw
> at it.
Oops. Maybe I spoke too soon! OK.
Consider... "...stable enough to assume an external reality..".
You are a zombie. What is it about sensory data that suggests an external
world? The science you can do is the science of zombie sense data, not an
external world. Your hypotheses about an external world would be treated
as wild metaphysics by your zombie friends (none of which you cen ever be
aware of, for they are in this external world..., so there's another
problem :-) Very tricky stuff, this.
The only science you can do is "I hypohesise that when I activate this
nerve, that sense nerve and this one do <this>" You then publish in nature
and collect your prize. (Except the external world this assumes is not
there, from your perspective... life is grim for the zombie)
If I am to do more I must have a 'learning rule'. Who tells me the
learning rule? This is a rule of interpretation. That requires context.
Where does the context come from? There is none. That is the situation of
the zombie.
<snip>
>> ..but..
>> The sense data is separate and exquisitely ambiguous and we do
>> not look for sense data to verify scientific observations!
>> We look for perceptual/phenomenal data. Experiences.
>> Maybe this is yet another terminological issue. Sensing
>> is not perception.
>
> If the perception is less ambiguous that the sense data,
> that is a false certainty.
Less ambiguous means more information content. More discrimination. The
brain accesses the external world directly, not only via sensing. A
mystery of non-local access = "hard problem" = we don't know everything.
We have to admit to this ignorance and accept that we don't know something
fundamental about the universe. BTW this means no magic, no ESP, no
"dualism" - just basic physics an explanatory mechanism that is right in
front of us that our 'received view' finds invisible.
It means that the rules of the appearance generation system
(phenomenality) are NOT the rules we generate using it (which is what we
traditionally use to do science). It doesn't mean the rules involved are
inaccessable. Or that the traditional rules of appearances are wrong. It
just means there are 2 sets of rules, not just one - and that they must
intimately co-exist. The new set of rules has not even been started yet.
Again I stress... no magic, no ESP, no dualism... only elimination of a
pervasive assumption: that the rules of the appearance generation system
(phenomenality generation) are the rules (scientific laws of appearance
expectation) constructed using it.
>
>> > Any machine which looks for regularities in sensory feeds
>> > does the same thing. Are you saying that such a machine
>> > could not find the regularities or that if it did find the
>> > regularities it would thereby be conscious?
>> I am saying the machine can find regularity in the sensory
>> feeds - easily. That is does so does not mean it is conscious.
>> It does not mean it has access to the external matural world.
>> ..and that is not what WE do....we find regularity in the
>> perceptual fields.
>> Looking for regularity in sensory data is totally different
>> process fro looking for regularity in a perceptual field.
>> Multiple sensory feeds can lead to the same perceptual field.
>> Multiple perceptual fields can arise out of the same sensory
>> feeds. No matter how weird it sounds our brains can map
>> sensory fields to the world outside...The sensory data is
>> intrinsically ambiguous and not about the external world,
>> but the world at the location of the transduction that
>> created the sensory measurent.
>> And an indirect transduction at that (a retinal cone protein
>> isomerisation by a red photon is not a 'red photon experience'
>> it is a protein-isomerisation experience - ie no experience
>> at all!)
>> I had no idea people were so confused about the distinction
>> between sensing and perception. I hope I am helping.
>
> It's an important distinction, but I think it is more complex
> than the dichotomy sensing/perception suggests. There are
> multiple stages from the interface between the environment
> and the sensory organ and the ultimate perception,
> and some processing of the signal occurs at each stage:
> nerve ganglia, the thalamus, various cortical levels.
> What we normally call "consciousness" seems
> to occur at the highest level of processing, but that may
> just be because that highest level is the dictator with
> ultimate control over, for example, what my
> hands are typing at the moment. Phenomena such as blindsight
> and the appearance of consciousness in other animals with
> less complex brains at least
> raise the possibility that some sort of consciousness/perception
> may be ocuuring at lower levels of processing in the
> nervous system, maybe all the way down to
> ganglia. Of course, ganglia don't come up with scientific
> theories, but this may not be because they are unable to
> perceive but simply because they lack sufficient
> processing power. Dogs don't come up with scientific theories
> either, and most people think that is because dogs are not
> smart enough rather than because they
> are not conscious. Similarly, our current machines may be able
> to "perceive" at a basic level, but they aren't smart enough to
> come up with scientific theories based on their perceptions.
>
> Stathis Papaioannou
>
All good stuff. In a very real way all behaviour is an experiment to test
a hypothesis of an expectation of appearances in a phenomenal scene. The
scientific act is very nearly hardwired into our systems. We have added a
whole bunch of procedural matters to make formal science possible and we
create communicable abstractions of expected appearances.
The empirical evidence is that certain emotional phenomenal scenes
(related to homeostasis) and probably smell/taste are generated by small
cohorts in the basal regions. Situational emotions in the amygdala. Other
more complex phenomenal scenes are localised to collections of whole
cortical regions.
There is no reason why animals should not have phenomenal scenes. Indeed I
predict bats have a primitive monochronme visual scene based on sonic
sensing. What animals don't have is the non-phenomenal capacity to
abstract and remember like us. Their behaviour could be thought of as
'proto-science' in the sense I describe above.
The main thing is that the creation of a phenomenal scene is an optional
subset of the whole package enabling intelligence. Knowledge is stored
devoid of phenomenal content until recall and usage. My knowledge of my
car's number plate was present all though the typing of this email, but
only...OFH964...now has it entered a phenomenal scene, first in my
language/visual imagination on recall, then reported out through my
fingers and back in my visual scene.
I don't think you're anywhere near the zombie you claim to be! :-)
Methinks I better get some exercise! Gonna go ride my bike for a while.
cheers
colin hales
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Everything List" group.
To post to this group, send email to everything-list.domain.name.hidden
To unsubscribe from this group, send email to everything-list-unsubscribe.domain.name.hidden
For more options, visit this group at
http://groups.google.com/group/everything-list?hl=en
-~----------~----~----~----~------~----~------~--~---
Received on Sat Nov 25 2006 - 16:37:05 PST