Hi Lewis I think not - I point out that there is a profound difference in the consequences between collapsing anchors and what you point to in the examples I mentioned and you graciously took the time to respond to. More specifically, collaspe anchors (originally called amusingly enough, "collapse knees" since when Bandler and I originally trained the whiz kids and other groups in the use of elementary anchoring, we had them anchor the two states to be collapsed on the knees of the client sitting in front of them preparatory to verifying the effectiveness and collapsing the (anchors on) the knees) is a procedure in which the agent of change either detects the natural occurence of or deliberately provokes the occurence of (preferably through behavior as opposed that damn instruction - Remember a time when ___________ (intensive definition of state being accessed)) state A (let's say, the problem state); state B (the "desired" state - please note that the selection of the "desired" state occurs only in the classic application never in the new code as this flaw of assigning the responsibility for such selection to the conscious mind - the part of the client least capable of making the selection - is corrected in the new code - see especially pages 212 - 222 plus relevant footnotes in Whispering. Thus the classic collapsing anchors is a pattern in which two distinct states of the client are accessed (present and desired), the efficacy of the anchors are checked and then the agent of change fires off both anchors simultaneously. Since our kinesthetic system (unlike our visual system) is incapable of entertaining simultaneously two states (physiologically defined), an integration occurs such that in the future whatever external triggers (just another class of anchors) accessed the old response (present state) will automatically access the new ("desired") state which is now collapsed with the original state (the present state above). There are any number of things one can do with anchoring and states - collapsing anchors (and their associated states is only one possibility. Note that in the context thus far developed, collapsing anchors is equivalent to integrating the two states of the client thereby anchored. Now, with this distinction cleanly available, I consider your remarks: 1. You wrote, "As soon as you identify the positive intention you have induced a significant state change which collapses anchors in the original context." In no way does the identification (whether consciously identified or simply (the preferred case, in my opinion) unconsciously verified through the involuntary signals that there is, in fact, a positive intention behind the behavior to be changed) imply "a state change which collapses anchors in the original context - it is not even clear in some cases that the original context has been identified let alone have the anchors in that context been identified - certainly, there is no doubt that the anchors in the original context that you refer to have NOT been identified, nor have them been tested for effectiveness and in no way have they been simultaneously fired. To make your case, please identify how the original context was identified, how the incompatible anchors were identified, how they were tested of efficacy and how specifically they were simultaneously triggered. BTW, I completely agree with you that as soon as the positive intention is verified by an involuntary signal(s)as you say, "The change is actually done at this step". This is an important issue to note - and one that is not well appreciated by many.  2. You wrote, "triple description ...entering each perspective and ascertaining the information therein (consciously or unconsciously) gives an opportunity to collapse anchors either at the point of entry or when re-entering what becomes an updated first position." The operational phrase in your statement is "gives an opportunity". Well, yes, of course, there is an opportunity - there are opportunities having coffee with a mate but this is hardly an argument that such collapsing anchors is a part of the pattern any more than drinking coffee with a mate. In fact, all the applications of triple description that I am familiar with specifically state that the client is to pass through a separator state when moving from one of the postiions of triple description to another explicitly in order to avoid contanimation -"contamination" in this phrase is roughly equivalent to a partial integration (or collapse of states. Again - to make your case, please respond to the questions raised in 1 above such as which anchors specifically were involved, how were they tested, how precisely were they simultaneously triggered...? 3. You wrote, "In this new state (the high-performance state)you step back into the problem context and without consciously attempting to do anything differently, notice what changes...this is where the collapse anchors occur.' Again, I find myself in disagreement with your representation. What happens in step 4 - the re-entry into the spatially located original state (the one selected in step 1 and visited in 1st position in step 2) - is that you associate the high performance state activated through the new code game with the auditory and visual stimuli of the context selected, as you correctly say, without consciously attempting to do anything differently. And here is the distinction - in classic collapsing of anchors, two distinct states in the client are accessed, anchored - the anchors are verified for effectiveness and then the two anchors (and their associated states) are fired simultaneously - with the resultant integration (collapse of anchors) occurring. In the new code format (page 240 in Whispering), there is no integration (or equivalently in this domain of discourse - I am not even certain that it is coherent to speak of integrating the state of the client with a set of auditory and visual stimuli that define the context in which the change will occur. This is better characterized as insisting on an association between the two - and is more or less equivalent to what was sometimes called future pacing - clearly designed to ensure that the differences secured during the change work would occur automatically (unconsciously) in the appropriate context. It is functionally equivalent to step 5 in standard 6 Step Reframing but there is no integration or collapse of two distinct states within the client him or herself. Again, to make your case, the usual questions await you. 4. You wrote (and thanks for pointing out that I missed responding to this point originally), "Now the second part of my statement which you omitted was that if you get the earliest triggers for the problem behaviour then attach your new state (however generated) then in theory this triggers the new state without "crossing swords" with the old one....fine in theory...yet I suspect (from my experience at any rate) that some degree of collapse anchors occurs anyway." Here you are describing the pattern called in the new code, stalking. Stalking is deliberately designed as a pattern in which the task of the client (with the sharp-eyed and eared support of the agent of change) is to detect the earlies possible signs of the onset of the state that s/he wishes NOT to experience - the context of application is perfectly general although the pattern was originally designed to deal with two sets of issues: state transitions that occur so rapidly the client does not have the ability to detect it is occurring until it has already happened (the state transition to the undesirable state) or the state transition to the undesirable state occurs so rapidly that it is below the threshold of the client's ability to detect it is occurring until the client discovers that s/he is already in the state s/he does NOT want to enter(the notorious frog soup paradigm). I have also found the stalking pattern (or actually a variant on it) to be essential to creating choice in epilepsy among other "organic" challenges. I think that this exchange is very useful and to my way of thinking, it points to a formidible challenge - a taxonomy of applications of anchors and/or a taxomony of operations that one can prefrom with anchors in pursuit of difference (change). I will open the conversation (that I hope will interest others enough to participate) with the proposal: that in the classic application of anchoring (above or pages 212-213 in Whispering), we have an anchoring and a collapsing of two anchors that correspond to two distinct states in the client - these are two entities of the same logical type (see pages 295 - 301 in Whispering). A state in a client and the auditory and visual stimuli that define a context of interest for the client are two entities of a different logical type. Thus, as a first approximation (to something that will, no doubt, have a much more refined representation), I propose that one explicit requirement for an appropriate collapsing of anchors (and their associated states) is that the two entities to be collapsed must be of the same logical type. Lewis, my apologies for running so long – however, I think that the issues under discussion are important in the extreme for the skillful and effective (even artistic) application of NLP patterning – anchoring, after all, is ubiquitous in our practice. Finally, in all the representations you offered in this posting, you spoke of a problem state as an integral part of the process. Since I appreciate your context of application as a physician, I understand your selection of terms. However, may we agree that these patterns are far too powerful and effective to restrict their application to problematic states and issues – these are simultaneously, patterning for optimization of functioning (an impossible and at the same time extremely useful fiction)  - a point well appreciated, for example, by Feldenkrais practitioners. Thanks for the quote from the Behaviorists - while I am not a fan of Behaviorism (as is apparent in Whispering), the point is well taken. All the best, John