Since in Heyting Arithmetic (HA) all atomic formulas are decidable, a Kripke model for HA may be regarded classically as a collection of classical structures for the language of arithmetic, partially ordered by the submodel relation. The obvious question is then: are these classical structures models of Peano Arithmetic (PA)? And dually: if a collection of models of PA, partially ordered by the submodel relation, is regarded as a Kripke model, is it a model of HA? Some partial answers to these questions were obtained in [6], [3], [1] and [2]. Here we present some results in the same direction, announced in [7]. In particular, it is proved that the classical structures at the nodes of a Kripke model of HA must be models of IΔ1 (PA- with induction for provably Δ1 formulas) and that the relation between these classical structures must be that of a Δ1-elementary submodel. MSC: 03F30, 03F55.
[1]
A. Troelstra.
Metamathematical investigation of intuitionistic arithmetic and analysis
,
1973
.
[2]
Samuel R. Buss,et al.
Intuitionistic Validity in T-Normal Kripke Structures
,
1993,
Ann. Pure Appl. Log..
[3]
Saul A. Kripke,et al.
Semantical Analysis of Intuitionistic Logic I
,
1965
.
[4]
C. Smoryński.
Nonstandard Models and Constructivity
,
1982
.
[5]
Albert Visser,et al.
Finite Kripke models of HA are locally PA
,
1986,
Notre Dame J. Formal Log..
[6]
C. Smorynski.
Applications of Kripke models
,
1973
.
[7]
S. Buss.
On Model Theory for Intuitionistic Bounded Arithmetic with Applications to Independence Results
,
1990
.