Lets say, this, you see an internal image which triggers a comment (ID) that comment where is it? Now as the internal image triggers that (ID) you also have a sensation (K). The interesting question here will be, do all of those responses occour at same time or do they occour at a small but measure intervalls? What most will do are to say that its a small interval like, V-A-K. Easy to asume its a digital between those since thats how we like to think. The - between the V-A symblize the analogue relation between those 2 and can be measured by the submodality shift that occours between V and A. So there isnt a digitalshift its also a analogue shift... They do occour simontaniosly. What a model does if I get this right from whispering are, it show us a simplified version of what goes on so we can asume to be doing the same as the person who do the thing. The strategy model then are a simplifed model to measure a 4 tuple. Its not whats really going on. If we are to replicate human behaviour we need some rules or models to show us what to do and how to do that. Most of times its gonna be a  v-k and the (ID) will be the metacommentary to what the person sees and feels. The the synthesia v-k occours and then the (ID) occours as a comment to the v-K. Seems like metastates right? Now, if we asume that then if to keep a valid strategy we need to have all 3 set of representations, which is visual, auditory and kino. Now lets think tripple descriptions, first, second and third. First and second are the v and k. What then becomes the third? Since aditory dosnt seem to be the optimal solution then the person needs to have a (K) check soemwhere. Or else they run the loop over and over again, commenting meta to the meta, which isnt anything else than a strategy that lack the basic criteria in NLP. Then as Grinder and Carmen states in whispering, do not metastates add anything new to NLP or is even related to NLP according to them. Since the metastates will be a strategy that runs without a strong kino check. And for that matter a exit. If you take what M Hall prosper using the criteria which NLP are based on then metastates dosnt work. If you however use the set of criteria M Hall set then soemthing else happens. Its all about more or less accepting a set of criteria. Those criteria makes it to a model partly or more. However everything a person does rarely enters a startegy or a model since its a simplified version and are made to teach you to get the same skill as the person has. If you using sensory based info then you dosnt use language. When you born you see, feel and hear things which means nada to you. You then construct a model based on the surrondings. Which the becomes solid reality. Our mind in 3D. Any model including NLP will have to face this, its like the programes in a PC, we need a betetr opertivsystem than windows but ppl are reluctant to change it since it requires, time and money to write new programes. The demand isnt that high to change to a new system even if a new system exists. We face the reality we construct and the one we are born into. Thats create the wast difference and sameness and relationships between whats old and whats new. To many live in a pavlovian world where when u smile they feel like you mocking you and they draw a knife and kills you. Did happen to a referee at world cup 1998, he was killed beacuse the person did feel he smiled to mock him. The refereee prolly did smile to make the other one feel comfertable, that and the lack of sensorybased eveidence killed him. But personally I think it was the knife and the guy who held it that had learned a model which wasnt based on trusting the fellow humans. And frankly, I do udnerstand violence, i do understand why people can go crazy and do things like this, thats one thing i have got from doing this NLP. Most problems comes from people that states we need to agree on this and in MY way. If you dosnt suit that then bogger off. I appriciate the dialogue between grinder and M hall and had a good time modeling what the responses had to be true for that dialogue to even concur as it has. For me it been a great time. People are mostly wrong. Even when they right. We model, we do, we create, we go further expanding new areas and transfer knowledge to others along the way. If you use the criteria "useful" then you are already at a metalevel since if you use representations to measure usefulness then you already have a good metalevel. Do we really need more than that? Dunno.. Continues... Criteria...does rules..dam does rules.. /Robert Sweden. It seems like any model that comes are a model of the persons head. Bandler and grinder toghter added a blend of those heads. Apart, they i think didnt do as they did toghter. It become, different, a taste of soemthing else.. More useful?