[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: on the nature of trust

On Thu, 12 Feb 1998, Tony Bartoletti wrote:

-> Ed,
-> I do believe that your definition of Trust (in the sense analogous to
-> Information Theory) is the most elegant and useful one, to repeat:
-> > Trust: "Trust is that which is essential to a communication channel but
-> >         which cannot be transferred from a source to a destination using
-> >         that channel"  
-> (Author! Author!)

Thank you.

-> As I have said earlier, it seems vaguely derivative of Kurt Godel's
-> foundation-shattering result, which I paraphrase poorly here:
-> "In any system of mathematics sufficiently expressive, one can produce
-> statements that are simultaneously True, and yet cannot be Proven to 
-> be True within the given system."

I am familiar with that but here we have a different beast. ;-) Quite
common as I will show.

(this exposition may better explain the def's machinery)

First, the definition of Trust I provided is NOT a statement to be proved
because it is NOT a derivation. It starts a new branch in Information
Theory, in which we go through the looking glass, so to say.

Second, you can't expect to prove first principles -- they are. If they
are coherent or not, if they adequately mingle in with other parts of
other theories that are used, if they provide useful results, if they
correctly reproduce known results, etc., all this constitutes the acid
test of the definition (like a photoplate that is revealed by acid attack)
in that context. It may be that you need to revisit/revise the whole
theory in order to allow the new definition to mingle in or maybe you can
just make second-order adjustments. 

Third, this is what I call a "shadow definition". It is a simple and
effective tool, often used. A "shadow definition" is a def that qualifies
the subject by negating a property. For example, a neutron "is an atomic
particle that has no charge", or, a prime "is a number that has no
divisors save itself and the unit", or zero "is a number that cannot be
divided by itself", etc. Of course, some concepts (like zero) accept
shadow and normal defs.

So, the trust def is quite an ordinary type, as ordinary as the def of a
prime number, regarding its wording logic.

-> So the question remains:  If I must Trust the given channel by some out-
-> of-band means, wherein will this Trust originate?

Good question but needs a slightly more precise wording (ie, using the
terminology as defined, as it will be clear below). Maybe some answers are
in my posting to Carl.

First, I would change it to: 

Q1: If I must Trust the Information transfered from source A to
    destination B, wherein will this Trust originate?
A1: From any source, including A, that can be chosen at will by B, in
    real-time (see def in cie.htm) and with any desired redundancy (see
    def in cie.htm).

Q2: How will it be transfered?
A2: By using at least one channel which is NOT used to transfer the
    Information that has to be Trusted.


Dr.rer.nat. E. Gerck                     egerck@novaware.cps.softex.br
    --- Meta-Certificate Group member, http://www.mcg.org.br ---