Of course it has to be a hypothesis because the evidence is not there. How can it be otherwise? I am glad that you are not defending the argument from physics put forward by some of your confreres. Everyone and every business works on the basis of probabilities in such circumstances. Perhaps a librarian can explain why they would continue to subscribe to a journal when all the content was available free of charge. The quotation from Raym Crow (whose work incidentally I admire) needs to be taken in the context of his model in the same piece. To repeat - this disaggregated model leaves almost no role for publishers, so adaptation means giving up the role which academics want us to continue with or say they say or so they act. Interestingly the OA journals as currently constructed do not have a role either. There is some place under certification for overlay journals but that is all. Anthony