Re: Computationalism vs. Comp

From: Bruno Marchal <>
Date: Wed, 31 Aug 2005 15:48:06 +0200

On 31 Aug 2005, at 08:26, Lee Corbin wrote:

> Bruno writes
>> wrote:
>>> I find an assumption of COMP far more tenuous than an assumption
>>> of a natural world
>> I respect this.
> I think that there has been a good deal of confusion between
> (I) computationalism: the doctrine that robots running classical
> programs can be conscious
> (II) Bruno's theories which build on this long-standing belief
> (computationalism) and which go much further.

Choosing words is not always simple. I prefer to reserve the most
common clear name "STRONG AI thesis" for the doctrine that robot/
programs can be conscious. It is stronger than the Turing-like
BEHAVIORIST MECH Thesis, according to which a robot/programs can
behave like if it were conscious.
Now comp is stronger than STRONG AI, which is stronger than BEH-MECH.

I prefer to keep comp for YD+CT+AR, and COMP for its translation in
the language of a lobian machine, a state which we are still a long
way from in the list (due to the hardness to go through the barrier
of mathematical logic).

> This confusion has not been helped at all by Bruno continuing to
> use the term "comp" indiscriminately for both computationalism
> (which is also basically "functionalism")

In practice all functionalist are computationalist, but they believe
in some knowable level of substitution. Comp add the necessary nuance
that the level cannot be known. Comp is weaker than all precise
functionalist theses of the literature.
To be clear: Functionnalisme implies comp, and comp makes
functionnalism false!!!!!!! Thus functionalism is inconsistent,
although comp is very near it, but more modest: we cannot know our
substitution level. Comp is betting on the truth but unprovability of
some weak form of functionalism.

> and his valiant attempts
> to derive his "comp" from computationalism (involving use of Gödel's
> Theorem, etc.)

I derive physics from comp. Cryptically physics is given by an
integral (measure) on incompleteness.

> It must be added that I have *never* --- since 1965 when I argued
> for (what I didn't know was called) computationalism against others
> in my high school.
> It must also be stressed that Turing's most famous essay embraced
> what is today called *computationalism* and which---basically---
> was called functionalism in the 1980's and 1990's.

Let us try to stick on the names which have already been chosen.

> The YD (Bruno's rather picturesque way of describing uploading)

Yes. It is no more than that. But this helps for making the UDA
reasoning more easy.

> has also been argued about---especially by cryonicists---for over
> twenty-five years. When I first became a acquainted with it, we
> all called it "downloading": the notion that one's consciousness
> could be downloaded into a piece of silicon, with all the advantages
> of speed, durability, and backup capabilities that this entails.


> In 1989 or so the people that I hang out with began to call this
> "uploading" instead. You'll have no trouble with Google finding
> all the thousands of emails and papers written about uploading.
> The name was changed when it was realized that "downloading" oneself
> into a small or large silicon device had many disadvantages over
> "uploading" one's self into distributed, possibly Solar System wide,
> communications nets.
> TO BE SURE: the main point of contention among people is still whether
> functionalism is true.

So, just remember comp is the statement that there is a level of
substitution where I survive classical digital uploading.
Functionalist never takes into account the fact that we cannot know
the level. So I reserve the term "comp" for the case we acknowledge
that ignorance. Functionalist theory can be correct by betting
correctly the right level, but for getting the physics from comp you
need to take explicitly into account our comp-ignorance.
It is many subtleties of that kind which make very useful the use of
the non-trivial logic of self-reference.

> Is it true, in other words, that "if it sounds
> like a duck, walks like a duck, and acts in every way like a duck,
> then
> it's a duck!"? We who say *yes* to computationalism and functionalism
> are not in the same camp, as Stephen Paul King points out, as a number
> of notable theorists like Roger Penrose, who believe in their bones
> that there has to be a connection between quantum mechanics and
> consciousness.
> On the contrary, people who dismiss functionalism (computationalism)
> will hopefully realize their mistake before long if (when) robots
> attain the same behavioral capabilities that humans have.

People will have artificial brain before, I would think.

> On the other
> hand, if this proves to be truly impossible without quantum
> computation,
> then we computationalists will have to admit that we were wrong.

Actually comp, in the sense I am talking since the beginning, is not
incompatible with us being quantum machine. The non cloning theorem
does not make it impossible for the Universal Dovetailer to generate
all my quantuml digital states and to emlulates the quantum
computational histories. It generates a exponential slowing down, but
the quantum first person generated cannot be aware of that slow-down.
So Hameroff's use of quantum tubules is not incompatible with comp.
Penrose use of gravition is indeed explicitly not computationalist,
but to my knowledge it is the only one model/theory having that "non-
comp" property. This does not help him for tackling the mind-body

Received on Wed Aug 31 2005 - 09:51:18 PDT

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:11 PST