Antoine van Gelder on the 3-cloud method.
Antoine van Gelder who is the Professor of Internal Medicine a the University of Pretoria in South Africa and also a Jonah’s Jonah (i.e. a very qualified TOC expert) has kinldy agreed that I can post the following note on the 3-cloud method t this blog. It is a long post but I'm sure you'll find it worth your time.
If you wish to look at the argument of the 3-cloud approach being “fundamentally flawed” it would be appropriate to first examine the basis of the “standard” approach of TOC analysis.
“Effect – cause – effect:” Those of us who received our Jonah and Jonah’s Jonah (JJ) training in the early to mid-90’s were exposed to a different phrase: “effect-cause- PREDICTED effect”. I do not know who decided to drop the word “predicted” from the phrase, but I believe something was lost by doing it. I have always retained the phrase in my teaching. This is because it describes with elegant four-word precision the method of “building downwards from an observed effect (E) to a cause by the following method: “Check for the existence of a possible (“speculated, or postulated”) cause by looking for an effect (PE: predicted effect) that MUST be present if the postulated cause is present in reality.” This method is called in science the hypothetico-deductive method. (I am not aware that this term was ever part of formal TOC-training.) In some strange way, this checking of causes by confirming their existence from a predicted effect became known in TOC as the DEDUCTIVE METHOD. And this is often contrasted with INDUCTIVE METHODS as if it is superior. Of course – the finding a hypothetical cause is nothing else than inductive reasoning!
However the confirmation that comes from arguing:
“If postulated cause C exists, then Effect (PE) must also exist;
PE is found to exist on checking;
Therefore C exists” is problematic for at least two reasons:(1) Karl Popper popularized the hypothetico-deductive method as a general scientific method http://en.wikipedia.org/wiki/Hypothetico-deductive_method but it should be noted that this method fundamentally depends on MANY instances of checking predicted effects in a SINCERE EFFORT TO FALSIFY the hypothesis. This is not part of the TOC teaching, and for the type of group analysis that is performed even in a full Johah course is not achievable in the type of time available. In science more is required than a single confirming instance of a prediction. A very, very unlikely result, if anything else but the postulated cause is present, must be predicted, and found to be true for it to “confirm” the presence of the cause. This would be a “strong” test.
(2) However, the problem with effect – cause - predicted effect analysis is even more fundamental:
The formal checking phase for confirming the cause is:
“If C exists then PE exists;
PE exists;
Therefore C exists.”
This is nothing else than an instance of “affirming the consequent” and is formally recognized as a fallacy. A formal (or logical) fallacy is a pattern of reasoning which is always wrong! (the conclusion may be “true” but it cannot be accepted because the reasoning that got you there is false)
From Wikipedia: http://en.wikipedia.org/wiki/Affirming_the_consequent:
“Arguments of the same form can sometimes seem superficially convincing, as in the following example:
If I have the flu, then I have a sore throat.
I have a sore throat.
Therefore, I have the flu.
But many illnesses cause sore throat, such as the common cold or strep throat. Thus this argument is weak at best.”Now, how could I (and others who have recognized this problem) continue to apply and to teach as JJs this fundamentally flawed method, until the multiple-cloud approach evolved? (Those that did not consider this problem of course went ahead and did it and wrote about it in serene oblivion!)
The reason why teaching and using the method was acceptable is that the “formal way” of constructing the CRT is GUIDED by the INTUITION (in the sense of hidden or tacit knowledge, as Eli Goldratt uses the term) of the people who learn to use the method. The induction (hypothetical cause) is not random. The predicted effect used to “validate” or “prove” the existence of the cause is not random. Both are “known” to exist by the group who build the CRT from UDEs to Root causes. The predicted effect test is a “strong” test that gives credence to accepting a single test as validation of the existence of the cause (such as finding the flu virus in the blood in the above example). The test is strong if there are very few other causes of the predicted effect (although this is in no way part of the training) The method is highly intuitive. This is why Eli Goldratt has said that this type of thinking is the most difficult of all thinking, and is why, if this is done by an “outside consultant” (even the Jonah’s Jonah it is immediately rejected by the “insiders”!So we seem to have two “invalid methods”! The methods are only invalid if we ignore the first rule of an analysis “gather a number of people who care about and have intuition of the subject matter . . .” As with so much of the TOC body of knowledge, results often depend on carrying out ALL the steps of a process.
Now, why did I go over to the multi-cloud approach?
One problem is that some users of TOC still are in the mode of thinking of a “Root cause” entity that leads to most of the UDEs and is there fore THE core problem. In the formal TOC teaching there was a step to look for “tension” between two entities, and then to develop these two entities as part of the core conflict cloud. Many analyses did not succeed to find this “conflict” as it again is highly intuitive and takes quite some experience to find. It is often omitted and the core conflict entity accepted. Now, if the core conflict entity represents only one side of the core conflict cloud, then the implemented solution could easily turn out to merely do the opposite. But if the two entities are in conflict the result is that in a short while a whole new set of UDEs appear. (Changing an organization from “centralization” to “decentralization” is a case in point)The fundamental observation is that if the core problem is something like “centralize <--> decentralize” then decisions made on the basis of trying to achieve a compromise will be made all over an organization. Rules of thumb (heuristics) will evolve to resolve everyday issues and many of them will lead to flawed decisions resulting in UDEs. The core conflict will be seen in the conflicts the people face who made the flawed decisions. If you can “see” the core conflict cloud in these conflicts, why do you need the CRT to find the core conflict cloud?
To argue that a method that uses tacit knowledge of those that use it in the decisions what to put into an analysis and what to leave out and how to weight them is empty at the present stage. The TOC Thinking Processes work. In Goldratt’s words: “It uses the method of the physical sciences to get at the intuition of the people involved.” Method AND intuition. The method cannot be judged without the value of the intuition. The cloud is a powerful way of liberationg tacit knowledge. The multicloud approach is a major advance.
Recent Comments