In previous eras philosophers have discussed the question whether or not we inhabit the best of all possible worlds. Leibniz, for instance, concluded we did: Ainsi on peut dire que de quelque maniere que Dieu auroit cre le monde, ... Dieu a choisi celuy qui est le plus parfait, c est a dire le plus simple en hypotheses et le plus riche en phenomenes ... (Discours de Metaphysique 6, 1686; Voltaire, by the way, ridiculized this idea in his famous Candide).
In traditional logic, possible worlds have been dealt with in a more unbiased or abstract, set-theoretical sense. They have been used simply to represent the way the world might be according to the information of agents, or the way they might want it to be given their desires. No arbitration between these worlds takes place, except, possibly, for the assumption that one possible world is the actual one. However, here as well, things have changed.
For instance, in modal and action logics and in game theory, reasoning and interpretation are placed back right in the context where they occur: in the context of the agents situation, her beliefs and her desires. The general assumption is that agents, besides deliberating about the question whether we inhabit the best of all possible worlds, also seek to optimize their environment so as to create or maintain the best possible world attainable to them. Here, optimization is the key notion by means of which rational behavior, including the linguistic behavior involved in information exchange, gets analyzed.
Let us illustrate this with a couple of examples. If we hear that it is just not true that Bernd is not going to sell his shares in MacroHard, then most probably he is going to. But maybe he is not in a position to do this, either because he doesn't own any, or because he is simply out of control. The right interpretation of what we hear depends on an optimal rendering of the utterance relative to the utterance situation, including what we know about the speaker's beliefs and intentions and those of Bernd.
Similarly, we can explain or motivate someone's actions on the stockmarket in terms of her expectations about the other players rational reactions to the recent fluctuation of the shares of, say, Quantification and Co. Notice that, in both examples the expected actions or interpretations of the other players depend on that of one s own, so that a logical best response depends on a deliberate assessment of the practical and epistemological situation. This is where dynamic and epistemic logic come in.
Optimization also proves to be a logical respons to socio-political matters. Think of two hypothetical countries, Taxonia and Fraudesia, which have exactly the same social-economical system, but a different practice. In the first, people generally pay taxes because they do not want to erode the social security system; besides, the chance of being caught if one doesn t is high. In the second, however, nobody does, because nobody else does, nobody wants to be a thief of his own wallet, and because the chance of getting caught is small. Thus described, both systems are equilibria ( steady states ) in both an intuitive sense, and in the formal sense of game theory.
Now think of a group of policy-makers of one of the two countries, who want to transform existing practice into that of the other. Obviously, this involves reasoning about the games people play, and the ways in which they play these games when the settings are changed. Ideally, logic is the instrument to analyze, motivate or denounce the plans which the policy-makers might come up with.