(Anti-)Lucas argument, Turing version

"I can find the Turing heel of any machine, so no machine can model me"

A machine can do that too:

Yd(z,x) ¬E <=> ¬(Yz(x,x)=0)
Yd(z,x) = fs(d,z)(x) = fg(z)(x)
Yh(x,y)=0 <=> fx(y) ¬E
Yd(h,x) = fg(h)(x)
Yh(g(h),g(h))=0 <=> fg(h)(g(h)) ¬E
fg(h)(g(h)) ¬E <=> ¬(Yh(g(h),g(h))=0)

The machine which computes the (primitive) recursive function g(z) calculates, for any machine Mp, including itself, the machine index and input g(p) at which Yp is not the halting function, so Mp does not solve the halting problem.

Adapted from Mechanism, Mentalism and Metamathematics - An Essay on Finitism, Judson C. Webb, D. Reidel Publ. Co. 1980, pp. 230-1.