Subject: decisions, decisions! (definition of faith)
Date: Wed, 31 Mar 1999 20:55:37 -0800
From: "Richard H. McCullough" <rhm@cdepot.net>
Organization: Richard H. McCullough, Knowledge Engineer
To: "Richard S. Latimer" <latimer1@att.net>
CC: "Harold L. Fields" <hlfields@volcano.net>

Faith is an attribute of the method of evaluating evidence about some
proposition, P.  Let's say
    decision:1 := person:2 do decide evidence:3 = decide_123
where evidence & decision have the values
    F -- evidence supports { P has false }       --  decision is { P has false }
    ? -- evidence does not support true or false --  decision is { neither }
    T -- evidence supports { P has true }        --  decision is { P has true }
and
    r_decide is relation with format="person:1; proposition:2; evidence:3; decision:4"

The RHM definition of faith is
    decide_123 has faith iff [person; P; ?; T]
The RSL definition of faith is
    decide_123 has faith iff [person; P; ?; T] or [person; P; F; T]

These are easier to grasp using a 2D diagram of decision vs. evidence
RHM decision   T  -  x  -
               ?  -  -  -
               F  -  -  -
evidence          F  ?  T

RSL decision  T  x  x  -
              ?  -  -  -
              F  -  -  -
evidence         F  ?  T

I hope you understand my pictures & formalisms.

I will leave to a later discussion the question of different kinds & strengths
of evidence.  For now, I'll just make a quick remark related to RSL's five
categories.  A first cut (analogous to F,?,T) is a two state description of
evidence
    G -- guess
    S -- sense data
The next more sophisticated description might use 4 states
      G -- guess
     GR -- guess plus logical reasoning
     SR -- sense data plus logical reasoning
      S -- sense data
to indicate the relative strength of the evidence.  To relate to the previous
description:
     G  <=>    ?
    GR  <=>    ?- or ?+  (leaning toward F or T)
    SR  <=>    F+ or T-  (not quite so certain)
     S  <=>    F  or T

Let me know if you think this view is useful.
--
Dick McCullough
mailto:rhm@cdepot.net
http://www2.cdepot.net/~rhm
knowledge := man do identify existent



Subject: re: Latimer faith & contradiction
Date: Thu, 01 Apr 1999 00:49:50 -0800
From: "Richard H. McCullough" <rhm@cdepot.net>
Organization: Richard H. McCullough, Knowledge Engineer
To: "Richard S. Latimer" <latimer1@att.net>,
     "Harold L. Fields" <hlfields@volcano.net>

I've got a complete model for RSL 5 stages.  It utilizes four types of
evidence/decision-methods:
    G -- guess
    S -- sense data
    R -- reason
    C -- ignore contradiction
Using these terms & RSL definition of faith, we have
    RSL name    proposition   evid./method  supports    decision
    "know"          p1            S or SR |=  T            T
    "imagine"       p2            G       |=  ?            ?
    "guess"         p2            GR      |=  ?+           ?+
    "belief"        p2            GR      |=  ?+           T
    "faith"         p2            GC      |=  ?+           T+

Incidentally, I said "T+" instead of "T" just to emphasize RSL's sense of 
total commitment to the decision. 
--
Dick McCullough
mailto:rhm@cdepot.net
http://www2.cdepot.net/~rhm
knowledge := man do identify existent

