Re: [s-d] About A Nomic

[Cassie Bayer <kisse.bnomic@xxxxxxxxx>]

2009/2/19 comex <comexk@xxxxxxxxx>

> Next, the Turing Test:
> {
> A Player is an Agent who is capable of passing the Turing Test,
> consents to said designation as a Player, and has become a Player in
> the manner proscribed in the Rules.
> }
>
> The issue is that Turing Tests are designed for computers, and perhaps
> can't be passed by humans.  But either we are interpreting this
> hyperliterally or not.  If we're not, then we're fine, because clearly
> the intent of the rule was that any Agent who cannot be distinguished
> from a human (including, trivially, a human) can become a Player.  If
> we are being hyperliteral, I propose that humans are machines, and are
> therefore generally capable of playing the role of a machine in a
> Turing Test.  As it is reasonable that the questioner in a Turing Test
> might give a final response of the form 'X is a human and Y is not',
> which would be incorrect no matter what in a test with two human
> responders, humans are also generally capable of passing Turing Tests.
>  (Depending on your definition of 'the Turing Test', similar arguments
> apply.)  Therefore, no matter how we interpret the Rules, humans are
> capable of passing Turing Tests.
>

The problem here is that humans are the control state for the Turing Test.
 The Turing Test holds that a machine need produce an error rate among a
jury assigning "male or female" that is equivalent to humans performing that
same test.

The problem with this is that humans are not able to even take the Turing
Test, we're simply able to establish the baseline.

The Turing Test was designed to indicate when a machine has sufficient
intelligence to be treated as human.  Humans are a special case and in
mathematical formal logic: "Humans are outside of the domain of the function
'the Turing Test' which maps entities to a 'true' or 'false'."

To compose the function "The Turing Test", we have f(x) = { g(x) >=
g("human") }, where g(x) = { the error rate as a result of playing the game
specified }

So, either the Turing Test for humans is tautological, or it's outside the
domain of the function.  It all depends on how f(x) is defined.

I would hold that the rule should be rewritten, as "is equivalent in
intelligence to a human".  This would explicitly make the statement a
tautology for a human.
_______________________________________________
spoon-discuss mailing list
spoon-discuss@xxxxxxxxx
http://lists.ellipsis.cx/mailman/listinfo/spoon-discuss