Triple Loop Learning

Diversity Management - Triple Loop Learning - by Robert L. Flood, Norma R. A. Romm, Bob Flood. ISBN 0471964492 .

"[Diversity Management provides] intellectual contribution to the now widely debated issue of managing diversity. It carefully blends theory and practice in order to provide substance to the debate on diversity management in the social and systems sciences. The focus on triple loop learning increases the fullness of learning about the diversity of issues and dilemmas faced. It brings together three main learning centres in one overall awareness so that the process is more reflexive and those involved can operate more intelligently and responsibly." (Amazon review)


An idea from SwieringaAndWierdsma

See TripleLoopLockUp, WhenToUsePatternForm.

See also YouCantLearnSomethingUntilYouAlreadyAlmostKnowIt (which in a way addresses the first two loops).


Single-loop learning assumes that problems and their solutions are close to each other in time and space. They usually aren't; much of the literature (e.g., PeterSenge, WilliamEdwardsDeming, SwieringaAndWierdsma, MichaelMcMaster?) point this out. When the problem and solution are close in time and space, then just state the solution as a rule. See ChangeOfSetting. Here's what to do.

Double-loop learning incorporates insights about why a solution works. You need the insight to appreciate the pattern. I think most good patterns fall into this level of learning. They involve tradeoffs between forces, and insight into how to resolve the tradeoffs. Alexander says that a pattern has a set of forces that can be considered together as a unit independent of distraction by other forces or patterns. Here's why this works.

TripleLoopLearning involves principles that go beyond insight. This is harder for me to describe. SwieringaAndWierdsma note that it relates to identity: who are we, what business should we be in, how do the ethics of the solution match with our ethics. I believe that if you solve a software problem with an organizational pattern, you're working at this level. I haven't seen many patterns at this level, but I'm guessing they'd be dynamite patterns. Here's why you want to be doing this. -- JimCoplien

It's not hard to find these triple-loop patterns. This site, due to its XP interest, is full of them. OnceAndOnlyOnce, DoTheSimplestThingThatCouldPossiblyWork, etc. The point of these is not to inform you what to do, but to inform the solutions which in turn tell you what to do. So, while you may solve the problem X using Y, you aren't done until Y is the simplest thing that could possible work. Then you get Z, the third loop. -- SunirShah


When information cannot be processed within the given knowledge system, the system has to be revised. This is what Argyris and Sch�n call "Single and double loop learning". Single loop learning is based on negative feedback (as used in cybernetics) and correction of deviation from the norm. Cybernetic rules allow signalization of changes and reactions to it. Learning is limited to operational actions that are allowed. Single loop learning has applications in quality- and effectivity problems. It works at the level of facts. Learning is focused on the question how we can reach an existing goal in the best possible way within existing norms. Double loop learning is directed to correction of "mistakes" based on defined norms. It implies an extra loop (Senge) reconsidering existing rules and their validity. TripleLoopLearning (also Multiple loop learning) is permanent learning. Not only the facts and the norms are considered, but the context. Triple loop learning works on perception and the permanent questioning of inconsistencies and incongruencies in the actiontheory of the organization.

As I perceive this, it implies (a.o.) working with different metaphors, "seeing" the limitations of our context of "seeing".

-- MartineDevos


MartineDevos describes TripleLoopLearning as permanent learning -- and that captures part of the thrust of the Diversity Management book which led me to post this pattern in the first place. (Diversity Management: Triple Loop Learning - by Robert L. Flood, Norma R. A. Romm, Bob Flood ISBN 0471964492 )

Flood and company's third loop is the victory of power which mandates that loop 1 and loop 2 (The Right Thing / The Right Way) are followed. Thus unless this loop, loop 3, is broken there can be no progress in further learning. I'm not sure if permanent learning in the sense of Martine's comment above, isn't what one has when loop 3 is in force. It seems to me that loops 1 through 3 form a paradigm in the Thomas Kuhn sense, and breaking loop 3 at least intermittently is a requirement to go beyond the current paradigm. Coloring beyond the lines, getting outta the box! -- RaySchneider

In HowBuildingsLearn Brand tries to extend the concept to buildings. His analogy: SingleLoopLearning? is minor fixes or adjustments, like using a thermostat to regulate temperature. DoubleLoopLearning? is major fixes or changes, like converting a sitting room into a guest bedroom. TripleLoopLearning is "learning to learn", for example putting in raised floor to make future changes (involving rewiring or cabling) easier.


Something similar to this was discussed by GregoryBateson in Steps to an Ecology of Mind. I believe he called them Learning 1, Learning 2, etc. and he thought that people could get past Learning 2, but that no-one was doing Learning 6. His notion was that you could not only learn things, but also better ways to learn things, better ways to learn ways to learn things, and so on. I haven't looked at this for at least 20 years, but my recollection is that he was talking more about individuals than organizations, and that he gave examples up to Learning 4 or so. -- MatthewWilbert


I just looked up Argyris and Schon's original description of double loop learning, and I don't think there can be a third loop - that looks like simple one-upsmanship by someone who didn't know. Argyris and Schon use single-loop learning for when outcomes and expectations are fed back to select between strategies. They use double-loop learning for when result sets and strategies are fed back to adjust norms, values and priorities. I don't think there is a think beyond norms, values and priorities to feed back to. There are only two levels. -- AlistairCockburn


The point about TripleLoopLearning in the Diversity Management context, is that it is the introduction of power to mandate the process. This mandate tends to lock up further attempts to change things, since, after all, we are already doing the right thing, the right way. I might add that all the single-loop and double loop stuff is already a form of taking well-known ideas and giving them a catchy name to suggest you've discovered something new. There is so much to know that this may well be a reasonable strategy, but it still smacks a little of one-upsmanship to use the term Alistair uses above. BTW Alistair, I'm finishing up reading Agile Software Development -- Good Book! Thanks also for putting me onto Dee Hock. I'm reading his Birth of the Chaordic Age because I ran into your citation of him in your book. Also see TripleLoopLockUp -- RaySchneider

(See BirthOfTheChaordicAge)


-- FrankSchophuizen?


CategoryEducation CategoryBook CategoryLearningMethods


EditText of this page (last edited September 30, 2014) or FindPage with title or text search