Re: Symbol Grounding Problem

On May 11, 5:23 pm, Neil W Rickert <rickert...@xxxxxxxxxx> wrote:
Don Geddis <d...@xxxxxxxxxx> writes:
Neil W Rickert <rickert...@xxxxxxxxxx> wrote on Thu, 10 May 2007:
A TM cannot run Windows.
Why not?
That should be obvious. A TM requires that all of the data be
on the tape before execution begins. But Windows is interactive,
regularly prompting the user for input.
You think this is a serious problem?

It's a problem for those who take the Turing machine model too
seriously (Penrose and Searle, for example).

Actually, it's a problem for those who wish to assert that mind states
reduce to impelmentation independent TM equivalent system states.
Penrose and Searle just take them at their word.

Surely you could just record all user
input for a period (an hour, a day, a year). And then replay that input
for a PC running Windows, or for a Turing Machine emulating a PC running
Windows, and you'll get the exact same output.

Try that with a robot. You would have the robot moving around, based
on yesterday's sensory data. It's going to screw up.

By the way, that's not the Systems Reply since it's relying upon
that which is not in the machine to assert failure.

This is a silly red herring.

I don't agree.

It's equivalent to you viewing Windows as
"not a computation".

I think we should recognize that Windows is not a computation.

Neil, this is a very good point. After Stevan's brief appearance I've
been thinking back at the past. Weren't you around when this was
discussed back then or are you new to this position?

Sure, it does lots of computations, all of which can be modelled
on Turing machines. But the overall system is not a computation.
The "Systems Reply" to Searle should be recognized as pointing
out that the overall system action is not computation (though it
contains computation).

That's not the Systems Reply. The Systems Reply asserts that while
Searle as the active agent in the Chinese Room may not understand
Chinese the System of Searle and the rules and the room understand
Chinese. The hypothetical situation is such that the room and it's
contents interact with the outside world appropriately so that the
system
interacts appropriately can't be the basis of the Systems Reply.

Which is true only in a very technical sense that isn't
relevant in this newsgroup.

I think it very relevant.

Very much so.

Your bland statement, "a TM cannot run Windows", is -- at the least -- highly
misleading, without further elaboration.

On the contrary, your attempt to deny this is what is misleading.

Yes, there is no provision for interaction within the definition of a
TM's
behavior aside from initial state on the tape and final state on the
tape
when the TM halts.

One might be able to divide the actions of Windows into small runs of
a TM where the "halt" is at the point of input from the world and the
input
is a recording onto the tape for initial state.

You would have great difficulty programming the unix operating system in
lisp, but not in C. Lisp presupposes a suitable virtual environment, while
C and assembly are capable of creating that environment.
You're actually wrong about this. A basic assembly language instruction
manual for some CPU also has zero access to low-level hardware facilities,
just the same as the ANSI Common Lisp specification does.

I'm not sure what you mean by "assembly language", but it sure is
different from what I mean.


.

Relevant Pages

  • Re: Privilege-escalation attacks on NT-based Windows are unfixable
    ... Whereas this is a systemic design flaw in the Windows API. ... >applications and if necessary fix them or to get Microsoft to fix the API. ... The OS does not require secure processes to open windows on the user's ... of the more naively-written services that interact with the desktop: ...
    (comp.security.misc)
  • Re: Privilege-escalation attacks on NT-based Windows are unfixable
    ... Whereas this is a systemic design flaw in the Windows API. ... >applications and if necessary fix them or to get Microsoft to fix the API. ... The OS does not require secure processes to open windows on the user's ... of the more naively-written services that interact with the desktop: ...
    (comp.os.ms-windows.nt.admin.security)
  • Re: web app that can interact with other applications on the server
    ... > Windows has the concept of a windows station. ... > designed to not interact with a user. ... > same winstation that the service is in. ...
    (microsoft.public.dotnet.framework.aspnet)
  • [NT] Flaw in Windows Message Handling through Utility Manager Could Enable Privilege Elevation
    ... to promote the most advanced vulnerability assessment solutions today. ... Microsoft Windows 2000 contains support for Accessibility options within ... or through the Accessibility Utility Manager. ... Processes running on Windows interact with the system and other processes ...
    (Securiteam)
  • Re: How to protect from LocalSystem privilege escaltion
    ... vulnerabilities in Windows NT/2000/XP itself. ... This is not a new issue, and Windows NT/2000/XP were designed with this very ... Setting this registry key may break services that were designed to interact ... > desktop an can therby be used to gain LocalSystem privileges. ...
    (microsoft.public.win2000.security)
Loading