Re: A request for information please.

Curt Welch wrote:
Stephen Harris <cyberguard-1048@xxxxxxxxx> wrote:
Curt Welch wrote:
Stephen Harris <cyberguard-1048@xxxxxxxxx> wrote:
Curt Welch wrote:
Stephen Harris <cyberguard-1048@xxxxxxxxx> wrote:
Curt Welch wrote:
I'm not interested in your define
consciousness game. I believe in my first person experience.

Ok, good. Lets talk about "first person experience". What exactly do
you think that is? Show me the data to explain what you are talking
about please.

It is what I experience and believe to be true. It is similar to what
I see and believe to exist. No two people see the same tree but people
are still able to come to a consensus about what the word "tree" refers
to. The tree is external and consciousness internal but they are both
experienced by me and my sense of self which is related enough to the
sense of consciousness which I experience when I wake from sleep.

I agree for the most part with Dennett's essay in the book "Sweet
Dreams" called "A Third-Person Approach to Consciousness". Discussing
consciousness requires a lot of typing and I'm no longer interested
in the question(s) you pose related to consciousness. I no longer
care much about the "Other Mind" problem.

I'm not required to show you data in the First person experience of
consciousness. People reach agreement about its existence by relating
to their own first person experience which is correlated by public
description and context.

Well, I think you have basically answered the question above in what you
write here.

I'm not sure why you added "believe to be true" however. If I happen to be
seeing things which I don't believe to be true, I'm still having a real
experience. So that seems to me to qualify for the idea of first person
experience.


Quoting again, CW: "What do you think "first person experience" is?"

>> It is what I experience and believe to be true. It is similar to what
>> I see and believe to exist.

That isn't really precise because I could experience a delusion. I meant
more the way I apprehend reality. Though it may not be how reality truly
is (like Plato and the Cave shadows), humans share a perceived consensus
reality which is as objective as it gets. I mean I experience myself and
that is real, I don't entertain philosophical doubts that I exist, nor
think I'm a brain in a vat or a fragment of some type of virtual reality
and seeing is the largest part of believing, not philosophical reason.

You notice that "see" has a double meaning, visual usually, and a mental
comprehension. Same with perceive. Who is it that sees and comprehends,
or undergoes an experience? To me it surely seems like a sense of self.
In one sense, experience is used as a verb, a human process.

In another sense, "experience" is a noun and it has meant an event which
has happened to a human. When a human overheats it might be described as
she experienced a heatstroke. When you car overheats it has not been
called an experience and described in terms of a human diagnosis. If
your cpu overheats and reboots, it is not a good reason to transfer
a human word and describe this as: it felt hot, so decided to rest.

To assert that a computer is not an it, so that it can feel words
which have been reserved for humans, or actually have experiences
which means there is experiencer rather than an abstract logical
identification as in "the right program instantiates a mind", it
takes more than word manipulation, which is another reason I tacked
on "believe to be true". The intuitions we derive from experiencing
reality are much more reliable than philosophical speculations about
what reality might be like.*

What's very true is that we make large assumptions that since humans are so
similar, our experiences must be similar, so anything we sense, the natural
first assumption is that other people probably experience things in a very
similar way.


Yes*.

It's also very easy to assume that things very different from us, like a
rock, or a computer, will have a very different experience.


Yes*.

What I can't agree with, is when this basic idea of what it's like to be
human, is assumed to be limited to humans, and that other things, such a
robots, couldn't have _any_ of the same sorts of experience. There's just
no evidence to support the idea that we are unique.

"same sorts of experience" means something different to me. A computer
undergoes a cause and effect event of reaching temperatures too high
for continued good functioning. This doesn't require the computer to
feel too hot, or to be having an experience; if you used experience in
a neutral sense (event) if would be ok, not implying an experiencer with a sense of self to sustain the experience.

But, suppose a computer has sentience, some sort of self. Then it could
have experiences at least somewhat like a human. I do not assume that
to be limited to humans. What I reject is the argument, that if I feel
hot -> that is equivalent in a meaningful sense to when the computer
overheats, so that therefore "we" now both feel hot, and the computer
or machine is having an experience. Or another analogy, if I feel stress
and an iron bar undergoes stress, then we both have the experience of
feeling stress so that we both have some consciousnes/awareness etc.
This is what I meant by using experience* neutrally, but then going on
to form a conclusion that employs first person subjectivity of the word*

So I thought a human level AI could be engineered. I mean a GM chess
program was engineered. But most people don't think Deep Blue has
consciousness or experiences in the usual sense which means an
experiencer of the experience. And that doesn't include defining
a bunch of dominoes set out to spell your name when the first one is
tipped to count as an experience although the result has a pattern.




The evidence only
suggests that we are unique directly in relation to the uniqueness of our
physical features. WE have neurons, our digital computers don't. We have
light sensors, the computers have light sensors. etc. We have a lot in
common with robots, and we have a lot different. This leaves us with only
one logical answer to our similarity to computers - we have some
difference, and we have some similarities.


I think there is one major similarity that is needed. A sense of self,
or Iness, that organizes or influences much of our behavior in terms
of evolutionary imperatives.

If you want to use the word "consciousness" to encapsulate everything that
it is like to be human, it's damn near impossible to not pick up some of
the features of the robot we share - like the ability to see a red apple
and know we are looking at a red apple. But yet, many here seem to assume

A fairly common definition of knowledge, but not perfect, is justified
true belief. We know a red apple when we see it. A program, so far,
can identify an object as a red object which doesn't mean it knows
anything. "Know" is another anthropocentric work when attributed to
computers. You are trying to steer this in a 'one can infer similar
internal structure or properties based on similar observed behaviors'.
But functionalism doesn't drive such a conclusion: F. says different
structures can produce similar outputs. That doesn't mean that the
structures don't differ in other ways besides the observed similar
outputs. So your point doesn't refer to somthing interior.

I think what motivates your argument is the idea that if all the external behaviors match up between robot and human, then one can
infer that other, supposed to exist, internal descriptions if they
exist, (such as consciousness) will also result or follow from all
the external behaviors matching up well enough. The people who
don't accept your motivation will probably say there needs to be
a volitional (akin to free will) property of sentience, _also_.

So I think your reasoning follows Behaviorism which has lost its
presumption of plausibility. Behaviorism used to be the dominant
paradigm of academia. It took a lot of doubt to displace B. from
its Psychology niche and insert the new paradigm of Cognitive _
I have better things to do with my time than re-examine B.
That choice is down there with arguing about the existence of God.


that our feature set holds nothing in common with robots. This is the
position that I feel can't be defended and the one that drives me up the
wall when people say that humans are conscious and robots and other
machines have no consciousness. It's a position which just can't be
defined. It's only a position which can be believed in by faith.


I am not sure what you mean by feature set, apparently more than a
robot can do more than "a human with a pencil, paper and eraser
following a simple rule requiring new ingenuity".

Humans and bonobos don't have the same feature set despite the
98.5% DNA similarity. Before computers, say 1925, plants weren't
considered to have consciousness, nor were rocks or machines even
complicated ones that performed tasks that would be considered
intelligent if humans were to perform them.

Why should that viewpoint be revised? What new theoretical insight
has emerged which plausibly asserts that machines have consciousness
or can feel and that this capacity is spread out in a spectrum with
the more consciouss machines determined by how many behaviors they
can duplicate of the human range of behaviors? There has been no
new physical evidence to substantiate such a theory.

There has been the Church-Turing Thesis which is well-received.
Several authors have modified the CTT beyond its original scope
which has been used to argue for Computationalism.

Strong AI is the conjecture which proceeds on faith. Where is
the historical precedent that a machine is considered to possess
consciousness? Golem witchcraft from biblical days.

Where is the demonstration that machines or metals or plastic possess
consciousness in a continuum (spectrum)? Where is the test for this?
You see it is not the duty of the people who don't believe that
machines have consciousness to provide a test to prove it doesn't. Because it is very hard to prove the non-existence of something which does not exist.

Examining your claim of believing by faith. Skepticism is fueled
by the subjective experience of awareness, and, there is no evidence
to accept your argument that robots have consciousness. Do you
somehow think that your philosophical claims are going to outweigh
agreed upon subjective impressions of ourselves as having consciousness
and doubting that plants and machines have consciousness?

It may be that robots have or can have consciousness. But so far
all you can say is that robots and humans share some behaviors.
You don't have a way of proving that they share other properties
from just judging from their behaviors. Thus, you will probably
have to make an assumption that behavior is the same thing as
consciousness, and that warrants the description of religious
advocation since that idea has already lost its credibility with
the mainstream as a scientifically valid claim.

Your request for data is part of the Second Person approach to
verify consciousness, not the first, which I used because I don't
care to defend or assume the burden of a Second Person verification.

If there is no way to perform second person verification, how can you
verify that you have it in yourself? Why isn't self verification the same
as second person verification?


I think therefore I am. I mentioned this before. There is no way
to prove that one does not exist inside a virtual matrix or as a
dream inside the mind of a god. The extent of First Person or
subjective verification is experiencing yourself. That does not
require evidence other than your own expression. Second Person
verification wants you to help establish your First Person
experience. The Second Person can never know exactly that which
you have experienced, but may be able to relate it to his own
experience. Third Person verification is how would a Martian
who did or did not have consciousness understand the concept of
consciousness as experienced by a human.


If I want to verify I have two hands, I have to turn my sensory system
towards my hands. If I want to verify you have two hands, I have to turn
my sensory system towards your hands. It's the same procedure whether I'm
testing my hands, or your hands. If I want to test if you have some
property which we call consciousness, I have to turn my sensory system
towards you. If I want to test if I a property of consciousness, I have to
direct my sensory system towards me. I don't really see that there's an
important difference between whether I'm verifying the condition of myself,
or verifying the connection of you.


You experience or see my hands, your hands, and your consciousness.
You can see my hands but you can't experience my consciousness.
You do something like infer my consciousness from my behavior and body
and that is not the same thing.

There is something called the Other Mind problem. People with autism
have difficulty in granting a mind or personality to other people
so they have a lot of trouble in participating in role-playing games.
But unless very disadvantaged, most autists have a sense of themselves.


main many difference, is that I have neurons in my had which I use as part
of my sensory system. And they are already connected to other neurons in
my head - so they are the instrument I'm using to verify that my other
neurons are working correctly. I don't have an easy way to connect them to
the neurons in your head so I can verity your neurons are working
correctly. So the problem of me verifying our neurons are working
correctly becomes a difficult one for practical reasons, but not a a
difficult one for philosophical reasons.

Now, the above idea is all logical if you are physicalist. If you believe
in the possibility of some form of dualism, then we run into the
philosophical problem.

I'm not interested in this. I am content to discuss issues with those
who already believe that the convenience of terms like mind is useful.

I have no problem with the word mind. I have HUGE problems with people who
believe that only humans have them. There is nothing about our mind, that
doesn't seem to exist in some forms in our current robots, and in our
current chess programs, and in all our current computers.


Well, since you've been asking me for a test, I assume you don't have
one and thus have no evidence. Other than what you may claim as
philosophical evidence. The problem is that there are a lot of
conflicting philosophies, I mean words that end in "ism". The "isms"
often conflict and the Philosophers within any given ism often disagree
with each other. For instance do you believe in Panpsychism where
everything in the universe is pervaded by consciousness? Or maybe not
plants and rocks and screwdrivers are conscious, but most animals and
machines as complex as thermostats? Where is your philosophy which
says how consciousness/mind got reserved for

>our current robots, and
>in our current chess programs, and in all our current computers.

and old dogs, parrots etc. Actually I want to know how the physical
basis for some of these to have consciousness but hammers and nails don't. I gave you my possible physical reason, emergence.

How does the abstract realm of philosophy reach out and influence
causal physical events?

We assume humans have minds because of our subjective experience of
them which is confirmed by mutual agreement with others. If we didn't
have a subjective experience of mind, then we might very well doubt
reports made by other who said they have them; like sighting UFOs.

So if people are to believe computers have minds, one needs a more
convincing argument than Panpsychism or a variant thereof. Minds
actually refer or require a sense of self or volition. Where is
the mind of a computer stored when it is turned off, if it is
physical like we assume the human -> brain -> mind to be. Since
the contents of ram are erased, does mind start over each time a
computer is booted up with the only stable part the hard drive?


To me, the word mind simply means "brain" - or maybe, the operation of the
brain which we can sense but that others around us can't sense. This is
how most people seem to use the word. If that is what they mean, I have no
problem with it.

Please outline the experiment I could perform on a human to find out if
that human was having this thing you call first person experience. Or
outline an experiment I can do on myself to find out if I have this
thing you call first person experience.

It is the sense of your existence. If you don't experience yourself
then I am content for you to believe you don't exist. Have I missed
something? Where is the requirement for personal experience to have
to be validated by another?

The reason I'm asking is not because I'm confused about what it is for
humans. It's because I'm pissed off at the people who pretend they know
things about the minds of non-humans when they haven't even stopped to try
and figure out why humans have minds and what a human mind is. They will
call me an idiot for suggesting that our machines already have minds, but
yet they have no evidence to suggest what the machine has or doesn't have.

You are presenting the new idea so you bear the burden of proof. And it is not like you are presenting your idea with evidence, but your are
presenting a philosophical argument and they are notoriously suspect.

They have no physical definition of what the human mind is, or where it
comes from, they have no way to test for it, in human, or non-humans, but
yet they pretend that it's obvious that our current computers don't have
any type of mind and that anyone such as I, who suggests they do, is just
an idiot.


If you have a new idea then you provide an experiment to validate it.
If you can't it is not the responsibility of others to discover such
an experiment. I think you misrepresent the situation. There are some
people who reject a computer with a mind. But the majority are agnostic
and say seeing is believing. You have a zero argument if you think that
you can present a premise and if others fail to disprove it, then that
counts as evidence proving it. Such things are an open question. I find
the CRA much more convincing than the rebuttals but that's plausibility.
Nor do I find Computationalism coherent. But that is not proof.

If all you know about the mind, is that it's the private behaviors we sense
happening which others can't sense - or the cause of those private

others can't sense our private, but they do sense their own private
and have come to agree that it is the same type of experience.

behaviors, then we can assume for many reasons that others humans have a
similar mind. But we can't assume anything about what other machines do,
or don't have, since their behavior is so different from ours (we can't
even have a discussion with them about these questions).


I agree with this. I think a little more evidence that computers have
a mind is needed before it becomes quite reasonable to suspect they have
a mind. Evidence that the brain is a simple computer is dwindling.

If you want to define "mind" as "human-level private behavior" then sure,
our current computers don't have that type of mind. We don't' have any
computers that can produce the typical range of thoughts and memories a
human will have. But when I suggested to people that were trying to argue
that computers don't have any sort of mind, or any sort of consciousness
that they were simply choosing to defined mind as human-level mind, they
refused to bite. They just run back to the basic position that only humans
have minds and machines have nothing like that.


I think you would have to explain how some things are eligible for the
spectrum of consciousness but others are not, such as rocks and nails.
Otherwise you would retreat to Panpsychism which claims everything is
conscious. I think that one needs to know how humans have consciousness
first before you can decide if other things also have consciousness.
Nature has quite a few things which mimic other things. In some ways
their external behavior matches. In some ways their internal functions
are similar and they might both have lungs or a nervous system.

I can look at my desk and see a computer. "I" know that "I" am looking
at a computer. Is that the first person experience you are talking
about?

Yes, either that or a frightfully good simulation.

How is this any different from a robot with image recognition software
that is able to look at the same view of a desk, and also known that it
is looking at a computer? It can know it's looking at a computer
without
You don't know the answer to this question? I don't that a computer can
know anything in the same sense as a human, I've never endorsed all
those anthropomorphic terms projected onto a computer/program.

And this is one of the foundations of my frustration. We as humans, have
made up a *** load of words to talk about human behavior. And, because we
made up the words for the purpose of talking about humans, it's a socially
accepted fact that those words should not be used to talk about non humans.
The words have an implied meaning of "human performing behavior". If you
use them to describe a car, you have misused the word. You have implied
you are talking about a human when you are not. This is all well and good.

At the same time, these words have been around for thousands of years. But
only recently, in the past 50 or so years, have we started to build
machines that have powers approaching the powers of the very interesting
and complex human machine. So only recently, have the machines started to
duplicate some of the same powers.


Isn't this another case that "powers" means behaviors and that you are
inferring from the behavior that all the properties which produced the
behavior are also transferred to the "new entity"? You don't know that
consciousness gets transferred or duplicated in the new entity unless
you assume that behavior is all there is to consciousness. You might
like that idea but it is a minority opinion considered dubious by many.

Now, for example, we talk about humans "knowing" things. This is one of
the many words that for thousands of years, could only be applied to humans
because no other machines came close to being able to duplicate the type of
behavior we call "knowing" in humans. But now we have lots of machines
that duplicate most of the behaviors in humans. I can ask Google to tell

This is an example of your criteria of knowing in terms of behavior.

the URL of web sites that have my name on it. It can tell me these things
because I knows the answer. But yet, we shy away from believing that the
Google computers actually "know" the same way we do. Though we might talk
as if they known things, many people still assume that they don't actually
work at all like humans.


Most people are going to consider it an unwarranted assumption.
You aren't providing evidence, you are providing a speculation.

But why is this? Is it because they don't work like humans? Is it because
what they are doing is fundamentally very different than what a human does
when a human answers a question? Or is it because we have been conditioned
by our language to assume that "knowing" is something that only a human can
do?


Actually, I think the root is that people think of themselves as "I's"
which know things and the computer is not informing a part of itself
which has an existence like a human's I'ness does. I think, feel, know, believe, experience etc. The difference is perhaps at an instinctual
level which distinguishes sentience/being/prey/predator and it/object
/thing/tool/fruit.

What I sense happening, is that people resist the idea that computers can
known like a human can known, because they have been conditioned to think
that way by our language. Our language has tons of words that only apply

That is true in part. But larger is there is no evidence, but some
unconvincing conflicting philosophical speculation. Many people think,
quite rightfully IMO that philosophy is not a good predictor of reality.

to human behavior. If we see the same behavior in a non-human, we are not
allowed to use the same label for it (by strong social convention). We say
that humans answer questions using their knowledge. They do it with the
their own free will and may choose not to answer. Computers on the other
hand are just running a fixed computer program that forces them to do what
they do. We say that don't have knowledge, but that instead, they are just
programmed to do what they do. They have no free will.

But to me, all this is just wrong. People that want to understand what a
human actually is, and what they share in common with our machines, so we
can understand what type of machine we are, need to stop letting themselves
by biased by the social conventions of our language. Just because it's
socially wrong to say a computer has knowledge, doesn't mean that what
computers do is different than what humans do.


Like I said language/tradition partially but mostly being an alive Iness

I have other reasons besides behavior to assume other humans have minds.
We share the same evolution which merges the range of our capabilities.
This is the simplest assumption.

Yes, I have no problem with that.

To claim some machine has consciousness
without human structure requires additional assumption of functionalism.

When I say machines are conscious, I currently don't mean they have a level
of awareness and a level of behavior that equals any human. What I'm

I know that. Your argument proceeds a lot like: people cannot very
well define consciousness. Actually, they can't refute your viewpoint
because of that. But that doesn't make your position stronger. To be
convincing, you need to define consciousness as something other than
behavior, people have already rejected that notion. You probably can't
do that, define consciousness in another manner or produce a test.

People will probably accept a program except for paying taxes or voting
which passes the Turing Test. They might consider it to be conscious.
I doubt if the consumers or the scientists will know that it's conscious
the people who hold passing the TT demonstrates consciousness, will
continue to think so, and the people who don't will still think the same
way, that it isn't conscious. TT is really not a test of consciousness
but an arbitrary standard, maybe the best available (neuroscience?).

trying to argue is that the things people point to in humans when they talk
about our consciousness, seems to already exist at some lower level, in our
computers. So if you point our ability to see a red apple as part of what
it means to be conscious, then I don't see how you can't understand that

I do understand it although I wouldn't use your terms because I think of
them as having additional connotations. What you leave out, is how a robot and a human both receiving sensory impression which allow them to
identify to another part of the system that this is an apple means that
both systems have consciousness? Again I think this is your behavioral
criteria assumption which many people reject. Next, your philosophical
justification lacks the physical mechanism explanation which builds
consciousness incrementally. It depends on behavioral criteria still.
There is a possible emergence explanation. I doubt that you are going
to be able to define consciousness so as to include some very simple
machines and exclude plants.

To me, your points are similar and I see myself making similar rejoinders. It takes too much typing and on a topic which people
write books and thousands of papers about. So I'm tired and
snipped the rest, except for this last part.

SH: A computer cannot have first person experience without some
major conjectures which Computationalism typically provides."

CW:
I agree that we are making major conjectures when we talk about computer
awareness. But at the same time, I believe we have to make even larger
conjectures to try and argue the position that computers don't have any
type of first person experience.

SH: People think of themselves as having selves with consciousness and
this is from their personal experience. They don't consider this a
conjecture. They extend this idea of consciousness to others, which
does become a conjecture. But this is not as big a conjecture as
extending the idea of consciousness to a robot. Now if the robot
were a perfect replicant like Rachel in Bladerunner, I think we would
assume she and Ford both owned consciousness. But apparently Rachel
was made from human DNA. I don't deny that two physically identical
or even close to identical, bodies of humans would have consciousness
rejecting Chalmer's zombie idea. But you are talking about a non-human
physical substrate. You can't assume that because physical behaviors
are close to identical that all other properties of the human substrate
have been transferred or created in the robot.

That requires a theory of consciousness that you haven't presented.
By which I mean, the idea of machines gathering consciousness along
a spectrum contingent upon acquiring more identical behaviors is a
premise, but not any kind of theory. It needs to go from the general
to the specific. Not from the specific (human) to the general, robots
and other agencies which exhibit human behavior. That is because
human behavior doesn't contain information about whether consciousness
has been transferred by logical argument leading to a conclusion. You
would have to assume/presuppose that behavior begets the presumption
of consciousness. Well, you can do that. But what that conjecture
does not do is minimize the conjecture that humans use to assign
consciousness to other humans which has a basis in their own belief
in their own consciousness. They also have behavior. You are trying
to use behavior alone as less conjectural to assign consciousness.
People will see this as less evidence thus more conjectural than
using the criteria of behavior plus using "it takes one to know one"
which is the standard humans now use. People are not going to agree
with you that using "it takes one to know one" adds more conjecture,
but rather that it is a better way to recognize reality than a philosophical approach which invokes behavior alone for its merit.

Regards,
Stephen
.