The discussion about the reality of causes on the cybernetics blog (http://www.asc-cybernetics.org/2010/?p=2893) continues to be fascinating.
My views on this are slowly forming. I think that 2nd order cybernetics tries to set itself up as a way of thinking that doesn't exclude any other ways of thinking. In reality it fails to achieve this because it carries an implicit ontological position which is inconsistent. People fall into the ontological trap and struggle to get out.
Beer's position was that 2nd order cybernetics is not necessary. The ontology of 1st order cybernetics was based around models. At its best, it extended scientific method with ways of making sense of the world based on a few principles (Ashby's Law is the main one). No fuss about observers, etc. But models are like masks: the person wearing it will be changed by it, as will those they engage with. But it depends on the person, and doing the right thing is always the bottom line.
What's the difference between the cybernetics of models and the cybernetics of observing systems? I think 2nd order needs a lot more terminology to describe the observing process. 1st order doesn't worry about the observing process. The terminology, like all terminology, enframes in ways that models don't, because models are merely representations of causal mechanisms.