Logical dynamics of some speech acts that affect obligations and preferences (Q1024131)

This paper describes a deontic preference language and logic DPL with formulas \(O_{(i,j)}\varphi\), to be read that it is obligatory for agent \(i\) with respect to authority \(j\) to see to it that \(\varphi\), and formulas \([pref]_i\varphi\), saying that \(i\) prefers it to be the case that \(\varphi\), or that all worlds that \(i\) considers at least as good as this world satisfy \(\varphi\). These are treated in a customary way. Following the lead of van Benthem and Liu's dynamic epistemic upgrade logic, DEUL, DPL is extended to the dynamic deontic preference logic DDPL by the addition of dynamic action modalities \([!_{(i,j)}\varphi]\) and \([\#_i\varphi]\), with \([!_{(i,j)}\varphi]\psi\) read to say that after a successful command by \(j\) to \(i\) to see to it that \(\varphi,\psi\) holds, and \([\#_i\varphi]\psi\) read that after an act causing \(i\) to prefer \(\varphi,\psi\) holds. Of particular interest are statements \([!_{(i,j)}\varphi]O_{(i,j)}\psi\), and even \([!_{(i,j)}\varphi]O_{(i,j)}\varphi\), representing the change in normative status of \(\psi\) (or \(\varphi)\) following a command to \(i\). DDPL is axiomatized and claimed to be sound and complete with respect to its semantics. Because the logic of \(O_{(i,j)}\varphi\) is essentially of type K, rather than D (or SDL), DDPL can accommodate deontic dilemmas without inconsistency, and it can avoid deontic explosion, but only in the case of a dilemma of the sort \(O_{(i,j)}\varphi \wedge O_{(i,k)}\neg\varphi\) when the authorities are different, when \(j\neq k\). The contrast between the modalities \(O_{(i,j)}\varphi\) and \([pref]_i\varphi\) and between their dynamic counterparts \([!_{(i,j)} \varphi]O_{(i,j)}\psi\) and \([\#_i\varphi]\psi\), is supposed to reflect Austin's distinction between illocutionary and perlocutionary speech acts, and DDPL is supposed to be a step toward a logical foundation for Austin's account in terms of conventional vs. causal effects. Philosophically, however, Austin's theory is dubious. The logic DDPL is far more interesting simply as a form of dynamic update deontic logic.