Whoa easy there Friday What or who is it that you perceive that you are defending Law of Requisite Variety from? It is clear that you wish to argue that there is value in including the Law of Requisite Variety (LRV) as a tool within NLP applications. There are several issues here: 1. Are any of the formulations offered in various part of this conversation adequate paraphrases of Ashby's original work? This is a excellent point - the three paraphrases (or actually interpretations) of Ashby's LRV in question are: "In any connected system, the component in the system with the widest range of variability will be the controlling element" (Whispering, page 276) "In any system, that component of the system which is most flexible will be in control of the outcome of the system." "In any system, all other things being equal, that component of the system with more options (i.e., variety in behavior) will be most likely to achieved a desirable outcome." Your suggestion (if I understand correctly is that these interpretations may not be paraphrases of Ashby's work. This is presumably what you intend to communicate with the comment you follow your own interpretation with; namely, "Not quite Ashby, but somewhat closer to that ideal, I like to think." You further comment that "The flaw, I fear, is in the paraphrases..." I agree fully with your point here. This is the point I wish to make in cleaning up the language by changing the term "paraphrase" to the term "interpretation". A paraphrase is a faithful mapping of a communication from its original terms into an isomorphic representation cast in new terms while an interpretation does not have to meet any of these stronger requirements. As you go on to point out correctly, Ashby's formulation applies to closed systems. However, technically, it is not simply that "that makes it no less applicable to open systems.  It's just that the proofs are harder" Yes, proofs are harder in open systems (infinite sets, for example) tham closed systems (finite sets, for example) however, to support your contention that this "makes it no less applicable to open systems" requires something more than a declaration. To the best of my knowledge, there is no proven (in the strict sense of meeting the requirements for rigor acceptable within the mathematical community) counterpart of Ashby's results for closed systems in open systems available. If you have access to such a proof, I would be most grateful for a reference. Further, there is a fascinating open question involved here. Ashby's LRV is a beautiful result of a study in abstract closed systems. What Bandler and I did in the early days of the development of NLP was to make an interpretation of this result (Ashby's LRV) and include in the presentation of NLP patterning at the meta level. The interpretation is clearly NOT a paraphrase but further for me the point of interest is whether it is legitimate to incorporate a result from the study of abstract formal systems into the arena of human behavior. Carmen and I argue in Whispering that the importation of well proven results in physics into the arena of human behavior (or the study of living organisms of all types) is NOT a legitimate mapping either intellectually or in practice. Indeed, the pattern Bandler and I coded under the name of Cause Effect semantic ill-formedness was our recognition that not only were such importation intellectually flawed but that they results in great pain and suffering among those who accepted uncritically such importations. A general study of how the mappings from studies and their results in the arena of abstract formal systems into human activities awaits some genuis. The closest thing referred to explicitly in Whispering is the work Chomsky did in mapping from Automata theory to the structure of human language (page 92). However, even here please note that the mapping is from studies in Automata theory (study of formal abstract machines) onto the strucure of human language NOT onto human behavior.  Fortunately, for this conversation, the point that we are both pursuing does not depend critically on the existence of a counterpart proof for open systems nor on the appropriateness of mapping from abstract formal systems onto behavior in the world of living organisms. You wish to argue for the retention in NLP application of the INTERPRETATION of Ashby's LRV as its has had for you specifically and for many other the salutary effects of inviting you to increase the choices available to achieve important outcomes. We agree. Indeed, this is precisely what we intended in our remarks terminating our  critique of the presentation of Ashby's LRV (Whispering, page 278/279), " Nevertheless, however these refinements (to the interpretation of Ashby's LRV)are to be made, there is much to be said in favor of any program of continuous learning (whatever its original motivation)  - the constant addition of new patterning to the set of patterns already mastered by the agent of change" Thus, your purpose seems entirely congruent with our recommendations. If you wish to argue for a stronger version or some elevated status for Ashby's LRV in NLP applications, then we invite you to respond with some precision to the specific critiques offered in Whispering on the pages cited: the issue of failure to respect logical levels in the application of Ashby's LRV and the issue about controlling as well as the brute fact that in the context of change, it is precisely the patient with the smallest number of choices that psychotherapists have traditionally found most difficult to change: catatonic schizophrenics and obsessive compulsives. All the best, John