Identifying Intelligence - Natural and Artificial Reasoning - page 30

QUERY > myinverseN || True
[ [ False True ] [ True False ] [ True True ] ]
QUERY > myinverseN || False
[ [ False False ] ]
QUERY > myinverseN && True
[ [ True True ] ]
QUERY > myinverseN && False
[ [ False False][False True ] [ True False ] ]
The principle of selecting the action is very simple and can be illustrated using the
Truth Table for 'implies' (see Table 3.1 ).
The requirement is to chose a condition that can potentially give two possible
results that will make the rule (last column ' A implies B' ) either True or False .
The first is to consider a positive example where A is a vowel . There are only two
possibilities given in lines 1 and 2 where:

' Aisavowel 'is True . E.g. 'A'

The first possibility is if ' A is a vowel' is ' True' . In this case there are two potential
results for ' Vowel implies Even', which depends upon the two possible truth-values
of the statement ' BisEven' given in lines 1 = True and 2 = False). Hence we need
to examine the negative case in which B is odd, since the rule will not be supported
if it turns out that ' A is a vowel' :

' Biseven 'is False (i.e. B is odd). E.g. #7
The two conditions { [' Aisavowel 'is True] overlap when [' Biseven 'is False
(i.e. B is odd)] }, and this only occurs when ' A implies B 'is False (Line 2) . So this
is the line that should be used to test the hypothesis where we have the pattern [ True
False ] for tests [ isvowel iseven ]. Thus 'A isvowel False '. In

the function ' matchresult ' the variable?10 in the example below can be either True
or False . Given a list of pairs of values (e.g. [ [?10 False ] [ True?10 ] ]), two truth
tests (e.g. [ isvowel iseven]), and a value (e.g. 'A'), the tests are both applied to the
value, and the resulting pair of results is checked for a match against each pair in
the list of possibilities. Only one of the possibilities needs to match for a result to be
True . To make clear how this matching is taking place, the alternatives are expanded
so that we have the following possibilities:

= True ' or '#7 iseven =

QUERY > matchresult [ [?10 False ] [ True?10 ]][ isvowel iseven] ' A '
True

Here it is True that '( isvowel 'A') is True'

QUERY > matchresult [ [?10 False ] [ True?10 ]][ isvowel iseven] ' B '
False

Here it is False that '( isvowel 'B') is True'

QUERY > matchresult [ [?10 False ] [ True?10 ] ] [isvowel iseven ] #4
False
Here it is False that '( iseven #4) is False'