Just finished Bart Kosko – Fuzzy Thinking (1993)
Excellent read, though Kosko does not seem to find favour with later non-classical logicians – many notes recorded (see bibliography). Focussed on machine AI and control systems – but many analogous applications apparent. Some points that stick instantly …..
(1) Reasons given for decisions / actions (eg to an analyst) are invariably logical positivist rationalisations of what in reality were much fuzzier rules governing actual behaviour (cf Argyris) Also, like Argyris implies, makes an explicit east vs west cultural distinction in logical positivist trap. (Over stated perhaps re Japanese consumer goods industry AI – topical at time of writing ?) See also my earlier (!!) (1991) conclusion in Managing Change and Flexibility – Attitudes and Organisational Culture – Chapter 4 – Blockages to Organisational Learning. Spooky.
(2) FCM (Fuzzy Cognitive Maps) look like excellent methodology to turn “mind maps” into fuzzy analysis tool. Could consider applying to this research effort itself ??
(3) Linguistics / natural language is the key to fuzzy rules (cf Robot Wisdom and Gellish)
(4) Fundamental Fuzzy Logic – Is it a bird, is it a plane ? – Yes. Decisions, including classifications, are not binary. Even if statistical answer, the probablities do not have to add up to unity !!! Being partly “both” (or one of any number of dimensions) can be 0%-100% on each orthogonal dimension independantly. (Contrary to Aristotle / Witgenstein etc : not logical positivist perspective – closer to eastern “mysticism”) Also very closely allied to the whole school of management thinking around “Paradoxes”.
What about a real process industry example :
- is it a pump, is it a heat exchanger ? Yes, it’s a thermosiphon.
Or an Information example :
Is it a document, is it data ? Yes, it’s XML in Motion.
Or more generally ….
Is it true that XXX ?, Well, Yes and No.
The clues are in the real world. This looks blindingly obvious – why model something approximately rationally, when it is precisely irrational ? You better believe it.
Reading James Gleick – Chaos (1987) comparing with Ian Stewart – Does God Play Dice (1989).
Mostly same material content so far. Copy temporarily stolen by my elder son Tom. Need to buy another copy anyway, to replace the one loaned by Leon, now covered in notes / doodles !!
Chaos and Fuzzy Logic stuff seems to recognise many applications in physics / physical systems / machines on the one hand and in populations / ecology / economics / politics on the other, but don’t see much reference to systems on the human organisation / business scale. Scalability aspect fundamental ? Surely I’m not the first in this field ?
All confirms reasons to follow up Walsham and Myers’s more qualitative / behavioural “ethnographic” IS approaches, analysis based on “what are these people doing here ?”. (See MISQ).
Also to research Heylighen’s work on associative modelling for search engines and PCP (Principia Cybernetica Project) via Links page. Heylighen also supports a primarily “process” view of the world, with all other physical or conceptual entities, phenomena and attributes being manifestations of “things happening”. A man after my own heart if ever there was one !!!! Seems logical to me if you are modelling “business”, then process is what it is all about after all.
Also points to need to formally understand “Systems Theory” paradigm from 1950′s / 1960′s and probably “holism / holistics” too.
Bibliography page added – under construction. Intend to use to capture notes / reference points.