I would suggest going to the source and reading at least one introductory article by Michael Hall before continuing this discussion. If anyone reads one of those articles and still equates a meta-state with 3rd position I will be amazed. Meta-states isnt a model using criteria which dervied from the first models. Either, need for more and/or different criteria for defining a model according to Meta-states. Even better, read one of the articles with the filter "this is just third position" and see what happens. Then ask yourself if this filter is the same as3rd position. One of the problems are that meta-state are a model of reframing. Anyhting we discuss will get distorted in a sense in meta-states. If you try to read as you state "this is just 3 position" you are not in that position. That slight shift places you outside third and therefore its not a valid test. The scope in 3 position will include both 1 and 2 position, now that pretty well defined. Anything else outside that will have a bigger set of representations but still be third position. This actually using structure and process which NLP is more about. If you add anything else the risk are to be content driven models and they have limited warranty. Content driven is ok to use if you know what and why they work as they do. Meta-states are a set of content driven information. The strategy you are writing are a strategy which makes anything you read as a telephone game. First to read this is just a third position and then to ask yourself if the filter is the same as 3 position. Doing things like that are to put implicit information into a more explicit consious mind. The examples in WITW with percepetual positions was at least for me a good example of defining those in ways understandable for the modeling process. /Robert