Math 557 Nov 7
Practice on \(\mathsf{PA}\) and Computability
The power of induction
Recall the axioms of \(\mathsf{PA}\):
- P1 \(\quad\) \(Sx \ne 0\)
- P2 \(\quad\) \(Sx=Sy \to x=y\)
- P3 \(\quad\) \(x+0 = x\)
- P4 \(\quad\) \(x+Sy =S(x+y)\)
- P5 \(\quad\) \(x \cdot 0 = 0\)
- P6 \(\quad\) \(x \cdot Sy = x \cdot y + x\)
- Ind\({}_\varphi\) \(\;\) \(\varphi(0) \, \wedge \, \forall x (\varphi(x) \to \varphi(Sx)) \to \forall x \; \varphi(x)\)
But we need induction to prove this, even if we throw in another axiom:
- P7 \(\quad\) \(\forall x \: (x \neq 0 \to \exists y (x = Sy))\)
The system (P1)-(P7) is referred to as Robinson’s \(\mathsf{Q}\)
Capturing basic arithmetic of \(\mathbb{N}\) in \(\mathsf{PA}^-\)
We saw in class that \(\mathsf{PA}^-\) proves all true \(\Sigma_1\)-statements about \(\mathbb{N}\). This is due to two facts:
- \(\Delta_0\)-formulas are absolute between structures and end-extensions thereof.
- Any model \(\mathcal{M}\) of \(\mathsf{PA}^-\) is an end-extension of (an isomorphic copy of) \(\mathbb{N}\).
To prove the second fact, we need to show two things:
- The mapping \(\iota: n \mapsto \underline{n}^{\mathcal{M}}\) is an embedding (so \(\iota(\mathbb{N})\) is a substructure of \(\mathcal{M}\) isomorphic to \(\mathbb{N}\)).
- Any element \(y \in M\) with \(y <^{\mathcal{M}} \underline{n}^{\mathcal{M}}\) is itself some \(\underline{m}^{\mathcal{M}}\).
The first item is established in the next exercise.
For the second item:
More on the limitations of \(\mathsf{PA}^-\)
(Is this fact provable in \(\mathsf{PA}\)?)
Number theoretic functions
Gödel’s Lemma
If you feel like it, try your hands at proving Gödel’s famous Lemma on the \(\beta\) function:
Hint: To code a sequence \(x_0,x_1,\ldots x_{k-1}\) of natural numbers, set \(m=b!\) where \(b = \max(k,x_0,x_1,\ldots x_{k-1})\). Verify the numbers \[ m+1, 2m+1, \ldots k \cdot m+1 \] are pairwise coprime. Now try to apply the Chinese Remainder Theorem.