Recursive Function uses One Minimization

From ProofWiki
Jump to navigation Jump to search

Theorem

Every recursive function can be obtained from the basic primitive recursive functions using:


Proof

Let $f: \N^k \to \N$ be any recursive function.

Consider the minimization operation on the $k + 2$-ary relation $\map \RR {n_1, n_2, \ldots, n_k, y}$:

$\mu y \mathrel \RR \tuple {n_1, n_2, \ldots, n_k, y}$

Consider the

From Minimization on Relation Equivalent to Minimization on Function, this is equivalent to:

$\map {\mu y} {\map {\overline {\operatorname{sgn} } } {\map {\chi_\RR} {n_1, n_2, \ldots, n_k, y} } = 0}$.


So we can rewrite the statement of Kleene's Normal Form Theorem as:

$(1) \quad \map f {n_1, n_2, \ldots, n_k} \approx \map U {\map {\mu z} {\map {\overline {\operatorname{sgn} } } {\map {\chi_\RR} {e, n_1, n_2, \ldots, n_k, z} } = 0} }$.

From the proof of that theorem, we have that $T_k$ is primitive recursive.

Hence from the definition of characteristic function, so is $\chi_{T_k}$.



We also know that $\overline {\operatorname{sgn} }$ is primitive recursive.

We also have by hypothesis that $U$ is primitive recursive.

Hence all of $\chi_{T_k}$, $\overline {\operatorname{sgn} }$ and $U$ can be defined without using minimization.

So the only minimization involved in obtaining the values of $f$ in $(1)$ is the one explicitly mentioned in $(1)$.

Hence the result.

$\blacksquare$