Previous page of Issue

Coherence from Incomplete Sequences (Part 1)

SERIES: Coherence from Incomplete Sequences
AUTHOR: Jim Schofield


This paper addresses a part of the methodology involved in Well-Informed Speculation, which is an essential process in Science to turn both extensive and careful observation with various isolated and extracted relations into a much wider attempt at Explanation – in other words, to develop a Theory. The paper is mostly concerned with a type of “fuzzy logic” as it appears in dichotomous trees, where DON’T KNOW answers are allowed in addition to the usual YES or NO responses within a tree of structured diagnostic questions. It demonstrates how, in spite of a series of such DON’T KNOW answers, surprisingly useful conclusions can still be extracted.

Indeed, spectrums of diagnoses are often sufficient for quite definitive conclusions, and even certainty can be possible, even when several such DON’T KNOWs are entered in response to questions, which would be expected to lead to confusion, yet instead can be shown to be irrelevant by later clearly inappropriate following questions, which prove that the current path being followed is quite obviously mistaken and can be abandoned..

Finally, the value of correcting loops, to home in on the Truth, are described as an established part of the processes of Abstraction.

1. The usual arguments against Theories is pinpointed as reasons for their ever present inadequacies, but shown to stem from purely formal ideas of Truth and Proof.

2. The example of “fuzzy logic” in dichotomous trees is addressed to reveal surprisingly reliable results from what seems to be a wholly damaging lack of answers to the diagnostic questions.

3. The phenomenon of Spectrums of Diagnoses being delivered from such trees with multiple DON’T KNOW answers is explained.

4. A special computer program incorporating these features was supplied to a biologist researching the possible taxonomy of Tardigrades, which assisted him in delivering the first comprehensive taxonomy many years ago.

5. Fianally, the necessity of iterative improvements even in such areas as Categorisation was described by which such diagnoses are regularly and successively improved was also described.

Read Paper (PDF)

Left click to open in browser window, right click to download.

Previous paper in series