Print

Print


On Fri, 26 Nov 2010 07:00:45 +0000, Simon Grant wrote 

> This does not seem to me to be evidence that fuzzy logic is uncontroversial. 
> My claim is only that it is controversial, that's all: let's leave it with
> agreeing that we do not, indeed, want to get into a discussion of fuzzy
> logic, Bayesian logic or any other related topic. 

Fuzzy logic (as set by Zadeh) **is** "controversial". Its an engineering
approach.  Its "novelty" was taking Łukasiewicz--- or, more properly,
his take on Aristotle’s  Syllogistic -- to engineering problems. I don't think 
Zadeh ever suggested anything more and was himself--- from personal
experience-- not too set on "it". Like much of engineering and especially
computer science little  new ground is broken but new approaches are
suggested. The great value of his work was to introduce philosophical logic
to engineers--- something that some want nothing of.
While "Bayesian logic" and "Fuzzy logic" have their own semantics and pile
different mathematics onto their sandwiches they have more in common than
any of their proponents would ever want to admit. From my perspective they
also make the same fundamental mistakes. I'll, however, refrain from an
exegesis given the lack of relevance of any of these paradigms to the current
topic.

> It might be more useful to separate the source of any uncertainty from the
> statement of how uncertain it is.

That's asking for too much. The source does not matter. All we have done is
to break down "uncertain" to two subjective and personal types:

1) Questionable. I have a date but I'm not sure--- or even that I suspect that
   the date is wrong.
2) Unknown. I don't have a date with sufficient precision yet. 198u, for
   example, means that I know it was in the 1980s but hope to, at some later
   time (state of knowledge),know.
and a data collection type
3) I've not yet collected the data. 198| (to use MARC syntax) means that I've
   not yet gotten around to collect the date (incomplete).

198| expresses that the date shall be updated to include its year at some point.
198u tells us that, at this time, we don't know the year.
(1985)? tell us that we think its 1985 but are not sure.

Pragmatically I would handle 198| and 198u as 1980 decade precision.
(1985)? says something else. Its not even saying with certainty that the
event took place in the 1980s.

We could be crazy and add grade of certainty and data-quality:

- a) Known to be correct (observed, documented etc.)
- b) Likely correct ( p> 50%)
- c) Possibly correct (Might be but not likely)
- d) Likely incorrect (The date is expected to be wrong p ~ 0)
- e) Unknown (certainty unknown).

(1985)? with the grade (a) is equivalent to 1985.

198u says that its known (a) that the date in the 1980s. The 'u' says we know
nothing more.

Going back to by example about the date of the Great Flood and the birthdate
of... 


--

 Edward C. Zimmermann, NONMONOTONIC LAB 
 Basis Systeme netzwerk, Munich Ges. des buergerl. Rechts 
 Office Leo (R&D): 
  Leopoldstrasse 53-55, D-80802 Munich, 
  Federal Republic of Germany  
 http://www.nonmonotonic.net 
 Umsatz-St-ID: DE130492967