Computing Beyond the Church-Turing Limit

Dershowitz and Gurevich claim to have proven the Church-Turing theorem starting from a set of 4 reasonable postulates. (Bulletin of Symbolic Logic, vol. 14, num. 3, Sept. 2008, pg. 299) But their postulate II assumes fixed vocabulary. Humans, however, change their vocabulary words (and concepts) over time. My Asa H artificial intelligence also changes its vocabulary. (Trans. of the Kansas Acad. of Sci., vol. 109, no. 3/4, pg 159, 2006, www.bioone.org/archive/0022- 8443/109/3/pdf/i0022-8443-109-3-159.pdf) Their postulate I excludes nondeterministic transitions between states. I don't know how often humans flip a coin but my Asa H does employ random transitions under certain circumstances. Perhaps humans and Asa H go beyond the Church-Turing limit. (Trans. Kansas Acad. of Sci., 108, 3/4, pg. 169, 2005)