25.12.2013 Views

On the methods of mechanical non-theorems (latest version)

On the methods of mechanical non-theorems (latest version)

On the methods of mechanical non-theorems (latest version)

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

1. If Γ has an automatic model <strong>the</strong>n to each φ ∈ Γ we can assign a finite concrete twosorted<br />

n-tape p-(multi-)automaton (W φ , h φ ) that recognises <strong>the</strong> definition <strong>of</strong> φ in <strong>the</strong> automatic<br />

model. Using proposition 2.7.1 we can merge <strong>the</strong>se into one finite two-sorted multiautomaton<br />

(W, h) where each component is equivalent to a (W φ , h φ ). The structure W is<br />

finite and clearly a model for Γ ′′ ∪ Γ ′ ∪ {∀q ∈ R[T φ (q)]}.<br />

2. If Γ ′′ ∪ Γ ′ ∪ {∀q ∈ R[T φ (q)]} has a finite model W <strong>the</strong>n W is a multi-automaton. To make<br />

W concrete we use a homomomorphism provided by lemma 2.2.2.<br />

qed<br />

2.8 Concluding remarks<br />

With proposition 2.4.6, corollary 2.4.9 and corollary 2.4.10 we have shown that in finite transitive<br />

automata reachability is first-order, we have shown that <strong>the</strong> class <strong>of</strong> transitive automata is finitely<br />

axiomatisable. Moreover we have shown that <strong>the</strong> finite and concrete representatives <strong>of</strong> <strong>the</strong> class are<br />

versatile enough to replace automata with fixed alphabets in J.R. Büchis procedure for deciding <strong>the</strong><br />

<strong>the</strong>ory <strong>of</strong> automatic structures. With proposition 2.6.1, corollary 2.6.4 and corollary 2.6.5 we have<br />

proven <strong>the</strong> same for PTPS automata.<br />

We have investigated <strong>the</strong> relationship between a PTPS automaton that recognises <strong>the</strong> interpretation<br />

<strong>of</strong> a formula with <strong>the</strong> PTPS automata that recognise <strong>the</strong> interpretation <strong>of</strong> its sub-formulae.<br />

With proposition 3.3.2 we have shown that this relationship is in a sense first-order. The proposition<br />

relies on first-order definability <strong>of</strong> various variants <strong>of</strong> reachability in PTPS automata.<br />

PTPS-automata can be used to construct a semi-desicion procedure that terminates on input<br />

on consistent sentences only, as follows.<br />

1. Let Γ be <strong>the</strong> set <strong>of</strong> sub-formulae <strong>of</strong> <strong>the</strong> input sentence.<br />

2. Let Γ ′′ ∪ Γ ′ be <strong>the</strong> definition <strong>of</strong> <strong>the</strong> PTPS-automaton in corollary 2.7.3.<br />

3. Search for a finite model for Γ ′′ ∪ Γ ′ ∪ {∀q ∈ R[T φ (q)]} and terminate if one is found.<br />

In step 3 we use <strong>the</strong> fact that <strong>the</strong> class <strong>of</strong> PTPS-automata is basic elementary to make <strong>the</strong> procedure<br />

terminate when <strong>the</strong> input sentence has a model definable in an automatic structure. This procedure<br />

terminates on input on at least those infinity axioms that are true in Presburger arithmetic.<br />

<strong>On</strong>e idea <strong>of</strong> transforming a first-order sentence into a first-order description <strong>of</strong> automata and<br />

<strong>the</strong>n to search for a finite model <strong>of</strong> <strong>the</strong> transformed sentence has been proposed by N. Peltier<br />

[Pel09]. To define reachability N. Peltiers transformation introduces <strong>the</strong> element-relation and lets<br />

<strong>the</strong> carrier-set consist <strong>of</strong> sets <strong>of</strong> states. In contrast, <strong>the</strong> transformation outlined in <strong>the</strong> present paper<br />

has a carrier-set whose members are symbols, some <strong>of</strong> which are also used as states. To define<br />

reachability we allow for some flexibility in <strong>the</strong> set <strong>of</strong> symbols. N. Peltier says <strong>of</strong> his transformation<br />

that, “The main interest <strong>of</strong> <strong>the</strong> present work is to prove that <strong>the</strong> translation is feasible from a<br />

<strong>the</strong>oretical point <strong>of</strong> view”. The same can be said <strong>of</strong> <strong>the</strong> transformation proposed in <strong>the</strong> present<br />

paper. To make search for satisfiable interpretations in automatic models remotely feasible from a<br />

practical point <strong>of</strong> view <strong>the</strong> present author proposes to use <strong>the</strong> atom-strucures <strong>of</strong> finite representable<br />

polyadic algebras. How to use PTPS-automata to compute such atom-structures is <strong>the</strong> subject <strong>of</strong><br />

<strong>the</strong> second paper.<br />

73

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!