[ExI] Fwd: Paper on "Detecting Qualia" presentation at 2015 MTA conference

Brent Allsop brent.allsop at canonizer.com
Mon Feb 2 04:18:30 UTC 2015


Hi Stathis,

On 2/1/2015 5:18 PM, Stathis Papaioannou wrote:
> On 2 February 2015 at 10:44, Brent Allsop <brent.allsop at canonizer.com> wrote:
>> Hi Stathis,
>>
>> On 2/1/2015 5:52 AM, Stathis Papaioannou wrote:
>>>
>>> Aspects of consciousness, or if you prefer of qualia, can certainly be
>>> investigated scientifically, and a large part of neuroscience and psychology
>>> is devoted to doing just this. However, it is impossible to investigate
>>> scientifically if someone actually has qualia and what those qualia are
>>> like.
>>>
>> When I say you believe this is not approachable via science, I am talking
>> about the latter, which you clearly state is not approachable via science.
>> In the latter you are making the falsifiable prediction that you cannot eff
>> the ineffable.
>>
>>>
>>> If you claim to be able to detect qualia then what test do you propose to
>>> use to decide whether CMOS sensors have "an intrinsic qualitative nature" or
>>> not?
>>>
>> The prediction is that if CMOS's behavior is the same as some quality (which
>> we have likely never experienced before) that we will be able to present it
>> to our augmented binding system in a way that will enable us to compare it's
>> quality to all the other qualities we have.   Before we do this, we will be
>> like Mary, and know everything about the behavior of CMOS.  But once we know
>> what our zombine information description of CMOS qualitatively represents,
>> we will also know, qualitatively, what CMOS is like.
> Can you give an example of how you would go about this?
One of many theoretical falsifiable possibilities, would be replace 
glutamate in the synapse with CMOS.  If you did this, and experienced 
the new quality of CMOS, you would then for the first time, experience 
the new qualia, and finally know, qualitatively, what CMOS was like.  
Then, like zombie Mary, who before walking out of the room, knew 
everything about how CMOS behaved, would finally know how to 
qualitatively interpret that zombie information representing everything 
about the CMOS quality.  And we/she would finally no longer be CMOS zombies.

>
>>> It's not problematic imagining that the qualia would vanish if the
>>> substitution were made with parts lacking the redness quality. What is
>>> problematic - and the entire point of the experiment - is that the qualia
>>> would vanish **without either the subject or the experimenters noticing that
>>> anything had changed**.
>>>
>> That explains our miss communication, then.  What I was trying to say, and
>> what this says you missed, is that the testable theoretical prediction is
>> that you will not be able to get or experience redness without presenting
>> glutamate (or replace glutamate with whatever your favorite theory predicts
>> is responsible for elemental redness) to the binding system of your mind.
> I understand this: we will assume that you need *real* glutamate to
> have the redness experience. So if we use ersatz glutamate, that
> functions just like real glutamate but isn't real glutamate, you
> should get normal behaviour but absent or different qualia. We could
> perhaps do this experiment by replacing normal glutamate with
> glutamate made from different isotopes such as C-14 and O-17. This
> ersatz glutamate will function chemically perfectly normally, but your
> claim is that normal function is not enough to reproduce the qualia,
> you need the actual substance. So what do you predict would happen if
> the natural glutamate were replaced with ersatz glutamate?

I don't understand why you are bringing this up.  The prediction is that 
you will be able to, through trial and error, find all possible 
necessary and sufficient detectable properties that enable you to 
reliably predict when someone is experiencing real redness, and when 
someone is not.  If you ever discover any detectable property that 
produces redness, that you didn't know had a redness quality, before, 
your previous theory will have been falsified and you must then simply 
alter your sets of necessary and sufficient detectable properties, to 
include the new property.  The same is true with all other physics.  
Glutamate is well defined, you can make high quality detectors of 
glutamate, that will only give a positive result, with real glutamate, 
and nothing else.  The falsifiable prediction is that detecting real 
redness will be the same as detecting real glutamate.  If ersatz 
glutamate has a redness quality, then  you include that in the set of 
possible detectable properties.  If altering glutamate, making it ersatz 
glutamate, alters the redness quality, then, either way, you still know 
exactly what has and what does not have a redness quality.


>
>> Only when you replace the entire binding system, with a binding system that
>> is interpreting zombie information representing redness, as if it was real
>> redness, will it behave the same.  So, it will be behaving the same, but the
>> qualitative subjective nature of it's behavior will have completely faded,
>> and be absent.  The system only behaves the way it does, because it contains
>> interpreting hardware that is properly interpreting the zombie information
>> as if it was the real thing. This is a form of the vanishing qualia case
>> David Chalmers predicts is possible, right?
> Chalmers says that this would lead to a partial zombie, someone who is
> blind but says he can see normally and behaves as if he can see
> normally. He stops short of saying this is absurd, but I think if you
> allow for the possibility of partial zombies the whole philosophical
> edifice crumbles.
>

No, the prediction is that as long as you have not replaced the binding 
neuron, nothing you present to it, will ever say and know something has 
a redness quality, without real redness.  In other words, without real 
glutamate, you will not be able to throw the switch, between the 
simluated glutamate, and the real thing, and reproduce the behavior 
saying the simulated glutamate is the same as the real thing.   So you 
will never get to the next level of replacing the binding neuron, 
because duplicating the "that is real redness" behavior will not be 
possible.  You will be able to skip that step, and replace the binding 
neuron which, by definition, has hardware translation that is 
interpreting the zombie information, as if it was real redness.  But, 
obviously, you can only think of this as behaving as if it had real 
redness, because, by definition, the zombie information does not have 
it.  Without the translation hardware, properly interpreting the zombie 
information, which does not have redness, as if it did, it will not 
reproduce the behavior.

Brent










More information about the extropy-chat mailing list