Information and Intelligence - Natural and Artificial Reasoning - page 21

Purpose

Insight Reason

Perceptions Concepts Heuristics Criteria

Problem Solution

Abstract Abduce Deduce Validate

Fig. 2.9 The intelligence process 4

Chose hypothesis(the general case):
All (male who has no brothers and sisters) are (the only son of a father)
Then given (from line a):
(I) is a (male who has no brothers or sisters)
1. Therefore:
(I) is (the only son of a father)
We infer (from lines b. and c.):
(This man's father) is (the son of a father)
And using 1:
(This man's father) is (I)
Therefore using a known relationship:
(this man) is (my son)

What is not described by this formal layout is why we might choose this particular
set of facts to make these particular steps as against the infinity of other possibilities.
We never considered the line of daughters, or the many other human relationships
that might have been chosen. There is nothing in the rules of deduction that offers
guidance to a useful conclusion.
To solve a problem using deduction requires direction ; deduction needs problem-
solving knowledge that limits the choices amongst the known facts and possible
hypotheses. Such problem solving knowledge provides a compass from which to
steer our course through a labyrinth of possible steps. This guiding knowledge is
known as a heuristic , and we should include it in our intelligent process (see Fig. 2.9 ).
We can now extend the model of the intelligent process to include:

• perceptions that identify the combination of features for a useful abstraction,

• concepts which are a set of generalisations that can be fitted to abstractions,

• heuristics which select the route through to a solution,

• criteria which provide the basis on which to accept viable hypothesis.

• Purpose that governs all the above.