#29 Adverse argument shown as a supporting one

v0.7
open
nobody
5
2004-05-25
2004-05-25
No

Let us have two predicates r1 and r2,
which relate e.g. persons to persons.
Let us assert as monotonic / forward:
(implies
(r1 ?X ?Y)
(r2 ?X ?Y))

Let AA and BB be two persons and let us assert as
monotonic :
(not (r2 AA BB)).

Then (not (r1 AA BB)) can be inferred.

Nonetheless let us assert as monotonic:
(r1 AA BB).

Then the assertion (not (r2 AA BB)) vanishes in the
browser,
instead we see (r2 AA BB) with a dark brown ball,
which means: "Truth: Unknown",
and (r2 AA BB) can no longer be inferred.

That is ok, but clicking "Arguments: 2" shows:
Assertion :
(r2 AA BB) in BaseKB

Arguments Supporting Assertion :

Asserted locally by Cyc administrator on May 25, 2004
at 10:41:22

Argument : Deduction #47851

(implies
(r1 ?X ?Y)
(r2 ?X ?Y)) in BaseKB
(r1 AA BB) in BaseKB

... as if (r2 AA BB) was asserted locally.

If we now unassert (r2 AA BB), we still have it as
deduced
(with a green ball, but monotonic
- a bit different from the "Key for Browser Icons"
document,
which explains this green color as "Deduced default true
GAF").

If we now assert
(not (r2 AA BB)) as default,
we see (r2 AA BB) (!) with the blue ball and two
arguments:

Asserted locally by Cyc administrator on May 25, 2004
at 10:57:36

Argument : Deduction #47851

(implies
(r1 ?X ?Y)
(r2 ?X ?Y)) in BaseKB
(r1 AA BB) in BaseKB

We get the same color blue and the textual information,
when we had asserted as default
(r2 AA BB) instead of (not (r2 AA BB))!

Discussion