Hello John 1. You wrote, "(remember my comment in another thread that the modeling as defined in Whispering is NOT the exclusive learning strategy, it just happens to define modeling ala NLP - as we suggest)" Now I got it. I've taken your modeling steps too seriously :-), having missed the point that it's just definition (and certainly a way to drive for). 2. You state, "I would say that there is grave methodological problems modeling from 3rd - modeling from 3rd implies a full set of f2 mappings and the methodological problem is that such a modeler is not model the model but in significant part him or herself..." Well yes, I fully understand your arguments and agree with them. The reason of my mentioning modeling from 3rd is just because other people (I mean the school of Robert Dilts) are doing so. And this is also the approach I was taught at the Master-Practitioner course. Your paradigm is still new for me... but it seems more promising. By the way, have you heard of any response to your book from Robert Dilts' side? I would be eager to see his opinion, especially about the issue with (neuro)logical levels (even more eager if his answer would be as funny as Michael Hall's ones have been ) Apropos, how do you evaluate his (Dilts') contribution to the modeling as in NLP? In his book, "Modeling with NLP", he develops some concrete strategies, mostly verbal though, using 3rd position... Before I read your book I thought that both 2nd and 3rd are appropriate and useful and that the decision whether to hang more in the 2nd or in the 3rd should be made ad hoc... And I still can't decide which approach (your 5-phase one or the Dilts' one) is generally more useful. 3. These were my words: "By the way, in the description, how should one attest that he's OK with unconscious uptake you mention some clients. I think the description should be refined because in many activities there's just no clients involved. Who would be the client if we modeled a tennis player?" You answered, "In the case of a champion tennis player, the clients would be aspiring tennis players who either wish to be competitive at this high level of performance or any tennis player who wished to significantly improve their game. I honestly couldn't understand the first portion of your statement in this quote." Quoting from the book (page 350, line 5), "...More specifically, the modeler demonstrates the ability to secure the same class of quality responses from clients that the model typically does in approximately the same time frame". Here we can see a presupposition that there are some clients which the model receives responses from. In aforementioned case of a tennis player there are none, right? The preceding sentence in the book, "More specifically, the performance criteria will be satisfied if and only if the modeler is able to reproduce the effects of the patterning modeled from the source.", still sounds good. 4. You write, "Content models require a certain acceptance of beliefs, values and presuppositions - none of which are required in the operational application of a formal model. This is also, of course, the basis of the critique that Carmen and I offer of the so-called presuppositions of NLP." As for presuppositions, let us consider the following case. Take a formal (and even most formal of all formal models) model which enables us to do X. Take a human being which has a limiting belief that X is impossible. How is he/she going to use the model? I would say that every model has an implicit presupposition that "X is possible", where X denotes the essence of the model. And there can be more complicated issues. For example, let X involve Y and Y involve Z. Now our person desires to learn X but, in turn, believes that Z is, say, bad. You state, "Leaving aside the important question of ethics, it seems very clear to me that NLP has spread like wildfire for the simple reason it works across languages, cultures, personal histories... precisely because it has no content - yet handles any content - this is the value of the syntactic or formal approach." Here I would like to share some of my speculations with you. Let us take the classic NLP anchoring format with client's conscious identification of the desired state (resources, ...) But this identification is already content - and not a form! One can argue here (after all, you state in the book that the form/content distinction is still pretty empirical), but if we consider the new code change format for comparison purposes, there's no content there at all. We can say that classic code formats are not really formal, but "polluted" with content (as well as characterlogical adjectives?) This "pollution" has one rather important implication - it gives NLP agents of change an illusion of control. For example, you ask the client about what he lacks in the situation of question. He answers, "Confidence" and before you request him/her to enter the corresponding resourceful situation you already have some expectations of how it might be (people use to take more upright poses when they are condifent etc.), don't you? Apparently, "confidence" (or what) is not really content but, strictly speaking, it is not a form anyway. The advantage of the "content pollution" is that we can utilize some portions of our general "community map". Sometimes it is just useful (and "map is not a territory" is only to remind us of the fact that these two can differ but not that they always differ). New Code approach takes away this "sometimes useful" degree of control and some people really miss it, heh. To say, the New Code is not that popular here, in Russia. Thank you for your answers, Oleg.