A note on the interpolation property in tense logic (Q1366764)

One can gain additional knowledge of the bewildering wealth of modal and tense logics by investigating whether these logics possess the interpolation property or not. In the paper under review this is done for so-called normal bimodal propositional logics. These are certain kinds of tense logics. The author calls them simply tense logics for short. In these logics time runs a linear course symmetrically into future and past. It is shown that such logics of unbounded depth do not possess the interpolation property. A tense logic has bounded depth if its models possess no infinite chains of possible worlds. As Bull and Segerberg point out [\textit{R. Bull} and \textit{K. Segerberg}, ``Basic modal logic'', in: D. Gabbay et al. (eds.), Handbook of philosophical logic, Vol. II, Synth. Libr. 165, 1-88 (1984; Zbl 0875.03045)], this is equivalent to the fact that for some \(n\) a certain formula \(B_n\) is valid in the logic. A tense logic with unbounded depth is a logic which does not have bounded depth. As the author points out, the lack of the interpolation property does not necessarily imply that these tense logics do not have reason. The author shows that two particular formulas \(\varphi_1\) and \(\varphi_2\) do not have an interpolant. He uses the same formulas as \textit{L. Maksimova} used in her paper ``Interpolation in modal infinite slice logics containing the logic K4'' [Sib. Adv. Math. 1, No. 3, 52-74 (1991); translation from Tr. Inst. Mat. 12, 72-91 (1989; Zbl 0729.03011)] to prove the corresponding result for monomodal logics. But her proof could not be simply copied, even though \(\varphi_1\) and \(\varphi_2\) are monomodal formulas. However the present proof can be carried over to give the result of Maksimova. Probably it can be shown that metatheorems of modal logic can be transferred to bimodal logic under certain conditions. In order to show that no interpolant exists, a model and two valuations \(V_1\) and \(V_2\) are defined such that, for any possible interpolant \(\psi, \varphi_1\rightarrow\psi\) is true under \(V_1\) but \(\psi\rightarrow\varphi_2\) is false under \(V_2\). By giving a bimodal logic of bounded depth for which the interpolation theorem holds, the author shows that the condition of unbounded depth is necessary. Where the condition is explicitly used, cannot be seen in the paper. It is needed in constructing the model mentioned above for which the author refers to another of his papers [``Tense logic without tense operators'', Math. Log. Q. 42, No. 2, 145-171 (1996; Zbl 0858.03019)]. In this paper the material needed cannot be found immediately but has to be inferred from Theorem 12(i). Therefore a reference to this theorem would have been useful, especially in order to better understand the role of unbounded depth. Nevertheless that part of the proof which is elaborated in the paper is very well presented. On the whole the paper under review shows that the interpolation property is no more a natural property in tense logic as it is in monomodal logic. This is certainly useful to know but not totally surprising.