|
|
|
|
|
We first exhibit a monotonic scientist that identifies . Let GN be an index for N. Let be an index for . Let Gj be an index for Lj that can be effectively obtained from j. Let denote an index for that can be effectively obtained from j and m. Let . Now let M be a scientist that, for each s |
|
|
|
|
|
|
|
|
. |
|
|
|
|
|
|
|
|
It is easy to verify that M is monotonic and identifies . |
|
|
|
|
|
|
|
|
Next we show that no weak-monotonic scientist identifies Suppose scientist Mj identifies Consider the language Lj. Clearly, because otherwise and Mj does not identify Lj. Let , hence . Suppose s is an extension of Tj[m] such that and Such a s exists as and Mj identifies Thus we have , which violates the weak-monotonicity property for identification of by Mj. Thus Mj fails to be weak monotonic. |
|
|
|
|
|
|
|
|
The next proposition shows that there are collections of languages that can be identified by weak-monotonic scientists but not identified by any monotonic scientists. |
|
|
|
|
|
|
|
|
5.57 Proposition . |
|
|
|
|
|
|
|
|
The above proposition (together with Proposition 5.54) implies the following corollary. |
|
|
|
|
|
|
|
|
(a) ![0112-016.gif](0112-016.GIF) |
|
|
|
|
|
|
|
|
(b) . |
|
|
|
|
|
|
|
|
Proof (Proposition 5.57): This argument is based on a technique due to Lange and Zeugmann [119]. |
|
|
|
|
|
|
|
|
For each m and n, we define: |
|
|
|
|